Genes and Alcohol Consumption: Studies with Mutant Mice

Abstract

In this chapter, we review the effects of global null mutant and overexpressing transgenic mouse lines on voluntary self-administration of alcohol. We examine approximately 200 publications pertaining to the effects of 155 mouse genes on alcohol consumption in different drinking models. The targeted genes vary in function and include neurotransmitter, ion channel, neuroimmune, and neuropeptide signaling systems. The alcohol self-administration models include operant conditioning, two- and four-bottle choice continuous and intermittent access, drinking in the dark limited access, chronic intermittent ethanol, and scheduled high alcohol consumption tests. Comparisons of different drinking models using the same mutant mice are potentially the most informative, and we will highlight those examples. More mutants have been tested for continuous two-bottle choice consumption than any other test; of the 137 mouse genes examined using this model, 97 (72%) altered drinking in at least one sex. Overall, the effects of genetic manipulations on alcohol drinking often depend on the sex of the mice, alcohol concentration and time of access, genetic background, as well as the drinking test.

Introduction

Alcohol use disorder (AUD) is a multifactorial disease and its risk factors are determined by the interplay of genetic and environmental factors, combined with neuroadaptations following acute and repeated alcohol exposure. Alcohol targets ion channels and signaling cascades, producing intoxication, anxiolysis, and a sense of reward. After prolonged, repeated exposure, alcohol-induced changes in gene expression and synaptic function are thought to contribute to the development tolerance, sensitization, and compulsive consumption and drug seeking.

More than a hundred genes have been shown to affect alcohol consumption and other alcohol-related behaviors in mouse models (Crabbe, Phillips, Harris, Arends, & Koob, 2006). Excessive alcohol consumption is a common model of addictive behavior, and animal models of voluntary self-administration are valuable for profiling genetic determinants of AUD (Green & Grahame, 2008). The preference to drink alcohol is a reliable measure that depends upon mouse genotype and has been consistent across laboratories despite variations of the drinking protocol used. Pre-clinical models, in conjunction with human genetic studies, may expose overlapping target genes and identify the most relevant drinking models and biological systems associated with AUD.

In this review, we focus on a single phenotype, voluntary alcohol self-administration, and summarize the global genetic manipulations in mice published to date on this behavior. Most of the studies used two-bottle choice (2BC) tests, where mice had a choice between water and ethanol and access was usually measured in continuous 24-hour periods. In some cases, 2BC access to ethanol was intermittent (e.g., every other day), which typically (Hwa et al., 2011; Melendez, 2011; Rosenwasser, Fixaris, Crabbe, Brooks, & Ascheid, 2013), but not always (Crabbe, Harkness, Spence, Huang, & Metten, 2012), results in higher ethanol intake compared to continuous access. A few studies used four-bottle choice (4BC) access, where mice have simultaneous access to water and three different concentrations of ethanol. Because rodents distribute their drinking across the circadian cycle and because limited access to ethanol tends to increase intake, restricted access during the dark cycle is often used to study periods of high consumption and to model binge-like drinking in humans (Thiele & Navarro, 2014). In the classic mouse drinking in the dark (DID) test, drinking session times begin a few hours after the start of the dark cycle and usually last 2-4 hours over a few days. High levels of ethanol drinking and pharmacologically relevant blood ethanol concentrations (BECs) are achieved using this model (Thiele & Navarro, 2014). The scheduled high access consumption (SHAC) test uses fluid restriction to promote drinking of a low ethanol concentration (Finn et al., 2005). This chronic drinking model can also produce high BECs. Fluid access is first restricted and then gradually relaxed until the effects of fluid limitation are minimized. In operant self-administration tests, mice are trained to self-administer quantities of ethanol that produce moderate to high BECs. Removal of access to alcohol followed by restored access transiently increases consumption in dependent mice, and the effects of mutant genes on this alcohol deprivation effect (ADE) (Rodd, Bell, Sable, Murphy, & McBride, 2004; Vengeliene, Bilbao, & Spanagel, 2014), a model of relapse drinking, are also presented. As shown in the tables, the effects of some mutants can depend on the drinking test used as well as the ethanol concentration, time of access, genetic background, and sex of the mice (Vanderlinden, Saba, Bennett, Hoffman, & Tabakoff, 2015).

There are important considerations regarding genetic engineering methods, including the potential alteration of genes other than the mutated gene and the influence of the background strain carrying the genotype. C57BL/6J (B6) mice are a high alcohol drinking strain, and as the tables in this chapter demonstrate, occupy a central role in voluntary drinking studies. This review focuses on global homozygous knockouts, although a few studies used hypofunctional or overexpressing transgenic lines. Strategies to reduce confounding and compensatory effects of null mutations include the use of knockin mice and brain regional or cell-specific knockouts, and a few of these studies are noted. In this review, we do not debate the genetic engineering methods used; instead, our aim is to provide a summary of homozygous null or overexpressing mutants and their role (or lack thereof) in alcohol consumption in mice. Current mouse gene and protein names from Uniprot (http://www.uniprot.org) are listed in the tables, which in some cases differ from the nomenclature used in the published studies. We first review the effects of mutant neurotransmitter receptor subunits on alcohol drinking in mice and then examine mutations in other ion channel receptors, cannabinoid and opioid receptors, neuropeptides, kinases/enzymes, and immune-related genes.

Neurotransmitter Systems

γ-aminobutyric acid (GABA)

Alcohol potentiates GABA_A receptor-mediated responses and enhances inhibitory neurotransmission, and some of the top candidate genes implicated in alcohol consumption code for specific GABA_A receptor subunits (Trudell, Messing, Mayfield, & Harris, 2014). Deletion of the α1 subunit decreased ethanol consumption in operant and 2BC tests (Blednov et al., 2003; June et al., 2007), and knockdown of α5 reduced drinking in male (Boehm, Ponomarev, et al., 2004) but not female mice (Stephens, Pistovcakova, Worthing, Atack, & Dawson, 2005) (Table 1). However, loss of α2 or β2 subunits had no effect on drinking (Blednov et al., 2003; Boehm, Ponomarev, et al., 2004). Decreased drinking was also observed in mice lacking δ or ρ1 subunits of GABA_A receptors (Blednov et al., 2014; Mihalek et al., 2001). In rats, local knockdown of the δ or α4 subunit in select regions of the nucleus accumbens reduced ethanol intake (Nie, Rewal, Gill, Ron, & Janak, 2011; Rewal et al., 2009). Deletion of glutamic acid decarboxylase, the principal enzyme for synthesis of GABA, had no effect on ethanol consumption if the mutation was expressed on a B6 background; however, consumption increased in Gad2 knockout mice on a B6 × 129/SvJ (N2) background (Blednov, Walker, Iyer, Homanics, & Harris, 2010). Knockin mice with ethanol-insensitive α1 subunits did not differ from control in ethanol intake (D. F. Werner et al., 2006), whereas α2 subunits were linked with alcohol consumption in some drinking models using a knockin strategy (Blednov, Borghese, et al., 2011). Furthermore, human genetic association studies have nominated the GABA_A receptor α2 subunit as a top candidate in alcohol-dependent individuals, reviewed in (Enoch, 2008). A recent meta-analysis also identified a strong association of α2 with alcohol dependence along with evidence for α6 and γ2 subunits (D. Li et al., 2014).

Table 1. Alcohol consumption in GABA mutant mice.

GeneKnockout/Knockin#	Background	Operant	2BC	DID	SHAC	References
γ-aminobutyric acid type A receptor (GABAAR) subunit α-1 (Gabra1)	Not specified	↓ (30-60 min), males/females		↓ (2 h) 1B, post-operant		(June et al., 2007)
	B6 × 129SvEv		↓ females			(Blednov et al., 2003)
Gabra1#	B6 × 129SvJ		—			(D. F. Werner et al., 2006)
GABAAR subunit α-2 (Gabra2)	B6 × 129SvEv		— males — females			(Boehm, Ponomarev, et al., 2004)
GABAAR subunit α-5 (Gabra5)	B6 × 129SvEv		↓ males — females			(Boehm, Ponomarev, et al., 2004
	B6 × 129SvEv	— (1 h) females	— females			(Stephens et al., 2005)
GABAAR subunit β-2 (Gabrb2)	B6 × 129SvEv		— females			(Blednov et al., 2003)
GABAAR subunit δ (Gabrd)	B6 × 129/Sv/SvJ		↓ males/females			(Mihalek et al., 2001)
GABAAR subunit ρ-1 (Gabrr1)	B6 × 129S4		↓ males — females — intermittent,males/females	— (2, 4 h), males/females		(Blednov et al., 2014)
Sodium- and chloride-dependent GABA transporter 1 (Slc6a1)	B6		—			(Cai et al., 2006)
Glutamic acid decarboxylase 2 (Gad2)	B6 B6 × 129/SvJ N1 B6 × 129/SvJ N2		— — ↑	— (3 h) 2BC — (2, 4 h) 1B	— (30 min)	(Blednov et al., 2010)

^#indicates gene knockin;

–, ↓, ↑: no significant difference, decreased, or increased ethanol intake and/or preference, respectively, in mutant vs. wild-type mice. Male mice were tested unless indicated otherwise. Ethanol intake in the two-bottle choice (2BC) tests was measured in 24-h sessions. Drinking session times for the other tests are indicated in parenthesis. DID, drinking in the dark; 1B, one bottle; SHAC, scheduled high alcohol concentration. Recommended mouse protein and gene (in italics) names are from Uniprot. B6 refers to C57BL/6J mice.

Glutamate

Ionotropic and some metabotropic glutamate receptors are known targets for alcohol action (D'Souza, 2015). Mice lacking Gria1 (glutamate receptor ionotropic AMPA 1), i.e., the GluR1 subunit, did not differ from wild-type in voluntary ethanol consumption, stress-induced drinking, or in the expression of ADE (Cowen, Schroff, Gass, Sprengel, & Spanagel, 2003) (Table 2). Mice lacking Gria3 (encoding the GluR3 subunit) did not differ from wild-type in the number of reinforced lever presses during operant ethanol self-administration, although cue-induced reinstatement was reduced (Sanchis-Segura et al., 2006). There were no genotype differences in voluntary ethanol consumption or preference at any concentration tested (2-16%) in a 2BC test. Both genotypes showed increased ethanol consumption after a deprivation period following 14 weeks of 16% ethanol exposure, but the increase was initially blunted in the knockout mice. Although self-administration was not different between wild-type and mice lacking GluR1 or GluR3, ethanol seeking behavior after operant administration and resumption of self-administration after alcohol deprivation were reduced in GluR3 knockout mice, suggesting a role for the GluR3 subunit in alcohol seeking and relapse.

Table 2. Alcohol consumption in glutamate mutant mice.

Gene Knockout/Overexpression*	Background	Operant	2BC	4BC	DID	SHAC	References
Glutamate receptor 1 (Gria1)	B6N		— — ADE — post-stress				(Cowen, Schroff, et al., 2003)
Glutamate receptor 3 (Gria3)	B6N	—	— ↓ ADE, day 1				(Sanchis-Segura et al., 2006)
Glutamate receptor ionotropic, NMDA 2A (Grin2a)	B6		—				(Boyce-Rustay & Holmes, 2006)
Metabotropic glutamate receptor 2, mGluR2 (Grm2)	CD1		↑ intake				(Zhou et al., 2013)
mGluR4 (Grm4)	CD1 × 129/SvJ		— males/females				(Blednov, Walker, Osterndorf-Kahanek, et al., 2004)
mGluR5 (Grm5)	B6 × 129/SvJ		— males/females	↓ females	— (2 h) 1B, females ↓ (3 h) 2BC, females	— (30 min), females	(Blednov & Harris, 2008)
	B6		↓				(Bird et al., 2008)
Excitatory amino acid transporter 1, GLAST, EAAT1 (Slc1a3)	B6		↓ males/females				(Karlsson et al., 2012)
Homer protein homolog 2 (Homer2)	Not specified		↓				(Szumlinski et al., 2003)
	B6 × 129Xi/SvJ		↓ 12%, males/females				(Szumlinski et al., 2005)
	B6 × 129Xi/SvJ				— (2 h)		(Lum et al., 2014)
Homer2b*	B6	↑ (21 min), NAc					(Szumlinski et al., 2008)
Neuronal pentraxin-2, NARP (Nptx2)	129Sv × B6		↓ intake, no escalation				(Ary et al., 2012)

^*indicates gene overexpression;

–, ↓, ↑: no significant difference, decreased, or increased ethanol intake and/or preference, respectively, in mutant vs. wild-type mice. Males were tested unless otherwise indicated. Ethanol intake in the two- and four-bottle choice (2BC, 4BC) tests was measured in continuous 24-h sessions. Drinking session times for the other tests are indicated in parenthesis. DID, drinking in the dark; SHAC, scheduled high alcohol consumption; 1B, one bottle; ADE, alcohol deprivation effect. NAc, nucleus accumbens. Recommended mouse protein and gene (in italics) names are from Uniprot. B6 refers to C57BL/6J mice.

Mice lacking Grm2, which encodes the metabotropic glutamate receptor 2, increased consumption and preference for high ethanol concentrations in a 2BC model using an escalation procedure in which the ethanol concentration was increased from 3-17% over 80 days (Zhou et al., 2013) (Table 2). However, Grm4 KO mice did not differ from wild-type in ethanol intake or preference in a 2BC test using 3-9% ethanol (Blednov, Walker, Osterndorf-Kahanek, & Harris, 2004). Grm5 knockout mice differed from wild-type in some drinking models (Blednov & Harris, 2008). There were no differences in ethanol (3-12%) intake and preference in male or female mice in a 2BC test. Because the effects of deleting Grm5 were somewhat greater in females, only females were tested in 4BC, 1-bottle and 2BC DID, and SHAC tests. Female Grm5 knockout mice showed lowered ethanol intake and preference in the 4BC and 2BC DID tests, but there were no genotype differences in the other tests (Blednov & Harris, 2008). Decreased ethanol intake and preference were also reported in male Grm5 KO mice consuming either 5 or 10% ethanol in a 2BC model (Bird, Kirchhoff, Djouma, & Lawrence, 2008).

The Homer family of postsynaptic scaffolding proteins modulates metabotropic glutamate receptors and is involved in addiction-related neuroplasticity. Homer2 knockout mice drank less ethanol than wild-type (Szumlinski, Toda, Middaugh, Worley, & Kalivas, 2003) and showed aversion to 12% ethanol, with total ethanol intake less than half of wild-type mice in a 2BC test (Szumlinski et al., 2005) (Table 2). Infusion of Homer 2 into the nucleus accumbens reversed the effects on ethanol preference and intake in knockout mice. Furthermore, Homer 2b overexpression in the nucleus accumbens increased operant self-administration of 6 and 12% alcohol in B6 mice (Szumlinski, Ary, Lominac, Klugmann, & Kippin, 2008). In contrast to continuous access, there was no difference in intake between Homer2 knockout and wild-type mice in a limited-access DID model using 20% ethanol (Lum, Campbell, Rostock, & Szumlinski, 2014). These authors reported that infusion of a metabotropic glutamate receptor 1 negative allosteric modulator into the nucleus accumbens reduced drinking in the DID test in control but not Homer2 knockout mice, indicating a role for metabotropic glutamate receptors scaffolded by Homer 2 proteins in alcohol consumption.

A family of neuronal pentraxins regulates AMPA receptor aggregation and synpatogenesis at the postsynaptic density. Neuronal activity regulated pentraxin (Narp), also known as Nptx2, is an immediate early gene that is induced by synaptic activity and produces an AMPA receptor binding protein that enhances AMPA receptor clustering and function. Narp knockout mice did not escalate their daily ethanol intake after repeated exposure, and total ethanol intake in these mice was significantly lower compared to wild-type after 10 days of testing (Ary et al., 2012) (Table 2). Further study showed that Narp deletion shifted intake and preference away from the high alcohol concentration (12%) with repeated alcohol exposure. Narp induction may thus be important for escalating alcohol consumption under free-choice conditions.

Chronic alcohol exposure and withdrawal produce a hyperexcitable or hyperglutamatergic state, and restoring the balance of glutamate responses could be relevant for treating dependence (Rao, Bell, Engleman, & Sari, 2015). Acamprosate (FDA approved to treat AUD) may work by inhibiting a hyperglutamatergic state in the brain and helping to balance inhibitory and excitatory transmission. In mouse models, acamprosate was effective in preventing the development of withdrawal (Blednov & Harris, 2008) and reducing free-choice ethanol intake (A. J. Brager, R. A. Prosser, & J. D. Glass, 2011; A. Brager, R. A. Prosser, & J. D. Glass, 2011; Spanagel et al., 2005). The anticonvulsant gabapentin is a calcium channel blocker that can also modulate glutamatergic (and GABAergic) activity and showed some promise in treating AUD in a clinical trial (Mason et al., 2014).

Dopamine

The mesolimbic dopamine pathway and its role in the rewarding properties of drugs of abuse has been a long-term focus for addiction research (van Huijstee & Mansvelder, 2014). Studies consistently show that ethanol self-administration in operant and 2BC tests is decreased in D₁- and D₂- receptor deficient mice (Delis et al., 2013; El-Ghundi et al., 1998; Palmer, Low, Grandy, & Phillips, 2003; Phillips et al., 1998; Risinger, Freeman, Rubinstein, Low, & Grandy, 2000; Thanos, Rivera, et al., 2005); however, deletion of the D₂ long receptor increased DID (Bulwa et al., 2011) (Table 3). Overexpression of D₂ transiently increased intake in D₂ receptor knockout mice and decreased intake in wild-type mice (Thanos, Rivera, et al., 2005). Effects of D₃ and D₄ dopamine receptor deletion are less well studied, but there is a report of substantially decreased alcohol consumption in 2BC and DID tests in D₃ knockout mice (Leggio et al., 2014). An example of the complex role of sex and genetic background is illustrated by the DAT knockout, which decreased drinking in females but not males on a B6 background and increased drinking in males but not females on a mixed genetic background (Table 3).

Table 3. Alcohol consumption in dopamine mutant mice.

Gene Knockout/Overexpression*	Background	Operant	2BC	DID	SIP	References
D1A dopamine receptor (Drd1)	B6		↓ limited/continuous access			(El-Ghundi et al., 1998)
D2 dopamine receptor (Drd2)	B6	↓ (23 h)				(Risinger et al., 2000)
	B6		↓ males/females — ethanol-exposed			(Palmer et al., 2003)
	B6		↓			(Thanos, Rivera, et al., 2005)
	B6		↓ males/females			(Phillips et al., 1998)
	B6		↓ ↑ chronic stress			(Delis et al., 2013)
D2 long receptor	B6			↑ (4 h), males/females		(Bulwa et al., 2011)
Drd2*	B6		↓ WT ↑ transient KO			(Thanos, Rivera, et al., 2005)
D3 dopamine receptor (Drd3)	B6	— (23 h)	—			(Boyce-Rustay & Risinger, 2003)
	B6		↓	↓ (4 h)		(Leggio et al., 2014)
D4 dopamine receptor (Drd4)	B6129/Ola × B6		—			(Falzone et al., 2002)
Sodium-dependent dopamine transporter, DAT (Slc6a3)	B6		— males ↓ females			(Savelieva, Caudle, Findlay, Caron, & Miller, 2002)
	B6				↓ (40 min), males/females	(Mittleman et al., 2011)
	B6 × 129Sv		↑ 24%, males — females			(Hall, Sora, & Uhl, 2003)
Dopamine β-hydroxylase (Dbh)	B6 × 129/SvEv		↓			(Weinshenker, Rust, Miller, & Palmiter, 2000)

^*indicates gene overexpression;

–, ↓, ↑: no significant difference, decreased, or increased ethanol intake and/or preference, respectively, in knockout (KO) or mice overexpressing D2* vs. wild-type (WT)/control mice. Male mice were tested unless indicated otherwise. Deletion of the long form of D₂ produces overexpression of the short relative to the long form. Ethanol intake in the two-bottle choice (2BC) tests was measured in 24-h sessions, unless otherwise indicated. Drinking session times for the other tests are indicated in parenthesis. DID, drinking in the dark; SIP, scheduled-induced polydipsia. Recommended mouse protein and gene (in italics) names are from Uniprot. B6 refers to C57BL/6J mice.

Serotonin

The serotonergic system has important roles in mood and impulse control and is implicated in the development and maintenance of alcohol and other drug abuse disorders (Bauer, Graham, Soares, & Nielsen, 2015). Ethanol intake and preference increased in a knockin mouse line (Tph2KI) expressing a hypofunctional variant of the rate-limiting serotonin (5-HT) synthesis enzyme, tryptophan hydroxylase 2 (Sachs, Salahi, & Caron, 2014) (Table 4). Increased ethanol self-administration has also been reported in mice lacking 5-HT_1B receptors (Bouwknecht et al., 2000; Crabbe et al., 1996; Risinger, Doan, & Vickrey, 1999). Knockout of 5-HT₆ receptors had no effect on 2BC drinking (Bonasera, Chu, Brennan, & Tecott, 2006). However, decreased intake was observed in B6 mice overexpressing 5-HT_3A receptors (Metz, Chynoweth, & Allan, 2006). Mice overexpressing 5-HT₃ receptors in the forebrain also drank less alcohol than non-transgenic mice in a 2BC test (Engel, Lyons, & Allan, 1998). The 5-HT transporter is the main means for regulating 5-HT levels in brain, and deletion of the transporter decreased ethanol intake in B6 mice (Boyce-Rustay et al., 2006; Kelai et al., 2003; Lamb & Daws, 2013).

Table 4. Alcohol consumption in serotonin mutant mice.

Gene Knockout/Knockin#/Overexpression*	Background	Operant	2BC	References
Tryptophan 5-hydroxylase 2 (Tph2)#	B6 × 129S6/SvEvTac		↑	(Sachs et al., 2014)
5-hydroxytryptamine receptor (5-HTR), 5-HTR1B (Htr1b)	Not specified	↑ (23 h) 10% phase 1, males/females		(Risinger et al., 1999)
	129/Sv-Ter		↑ males/females	(Crabbe et al., 1996)
	129/Sv		— males	(Bouwknecht et al., 2000)
5-HTR3A (Htr3a)*	B6DBA/2J		↓ —	(Metz et al., 2006)
5-HTR6 (Htr6)	B6		—	(Bonasera et al., 2006)
Sodium-dependent serotonin transporter (Slc6a4)	B6	↓ (90 min)		(Lamb & Daws, 2013)
	B6		↓	(Kelai et al., 2003)
	B6		↓ intake 7%, females	(Boyce-Rustay et al., 2006)

^*indicates gene overexpression;

–, ↓, ↑: no significant difference, decreased, or increased ethanol intake and/or preference, respectively, in mutant vs. wild-type mice. Male mice were tested unless indicated otherwise. Ethanol intake in the two-bottle choice (2BC) tests was measured in 24-h sessions. Drinking session times for the operant tests are indicated in parenthesis. Recommended mouse protein and gene (in italics) names are from Uniprot. B6 refers to C57BL/6J mice.

Although selective 5-HT reuptake inhibitors and the 5-HT₃ antagonist ondansetron were not promising as first-line treatments for AUD, they may provide some benefit in treating select populations of alcoholics. A review of pharmacotherapies for AUD in preclinical and clinical models can be found in (Zindel & Kranzler, 2014).

Adenosine

Adenosine signaling is implicated in drug addiction and ethanol-related behaviors in preclinical models (Nam, Bruner, & Choi, 2013). Deletion of the adenosine A₁ receptor did not alter operant responding for ethanol in B6 × 129 mice (Houchi, Persyn, Legastelois, & Naassila, 2013) (Table 5). Although male and female adenosine A_2A receptor (A2AR) knockout mice generated on a CD1 background drank more ethanol compared to wild-type in 2BC tests (Houchi et al., 2008; Naassila, Ledent, & Daoust, 2002), no differences were found between A2AR knockout and wild-type mice on a B6 background (Houchi et al., 2008).

Table 5. Alcohol consumption in adenosine knockout mice.

Gene Knockout	Background	Operant	2BC	References
Adenosine receptor A1 (Adora1)	B6 × 129	— (2 h)		(Houchi et al., 2013)
Adenosine receptor A2A (Adora2a)	CD1		↑ males/females	(Naassila et al., 2002)
	CD1B6		↑ males — males/females	(Houchi et al., 2008)
Equilibrative nucleoside transporter-1, ENT1 (Slc29a1)	B6 × 129X1/SvJ	↑		(J. Chen et al., 2010)
	B6 × 129X1/SvJ		↑	(Nam, Hinton, et al., 2013; Nam et al., 2011)
	B6 × 129X1/SvJ		↑	(Choi et al., 2004)
	B6 × 129X1/SvJ		↑ sex not specified	(M. R. Lee et al., 2013; Ruby et al., 2014)
	B6 × 129X1/SvJ		↑ females	(Ruby et al., 2011)

–, ↓, ↑: no significant difference, decreased, or increased ethanol intake and/or preference, respectively, in knockout vs. wild-type mice. Male mice were tested unless indicated otherwise. Ethanol intake in the operant and two-bottle choice (2BC) tests was measured in 30-min or 24-h sessions, respectively, unless otherwise indicated. Recommended mouse protein and gene (in italics) names are from Uniprot. B6 refers to C57BL/6J mice.

Type 1 equilibrative nucleoside transporter (ENT1) is an ethanol-sensitive adenosine transporter and a main regulator of adenosine levels in the brain. Table 5 shows that ENT1 knockout mice generated on a B6 × 129 background consume more ethanol compared to wild-type littermates in operant and 2BC drinking tests (J. Chen et al., 2010; Choi et al., 2004; M. R. Lee et al., 2013; Nam, Hinton, et al., 2013; Nam et al., 2011; Ruby et al., 2014; Ruby, Walker, An, Kim, & Choi, 2011). Thus, studies from multiple labs support a role for adenosine in controlling alcohol-seeking behavior.

Cannabinioids and Opioids

The endocannabinoid system is involved in brain reward signaling and drug-seeking behavior (Panagis, Mackey, & Vlachou, 2014). In several studies of different genetic backgrounds, male and female cannabinoid 1 receptor (CB1R) knockout mice showed reduced ethanol intake and/or preference for ethanol than wild-type mice (Hungund, Szakall, Adam, Basavarajappa, & Vadasz, 2003; Lallemand & de Witte, 2005; Naassila, Pierrefiche, Ledent, & Daoust, 2004; Poncelet, Maruani, Calassi, & Soubrie, 2003; Racz et al., 2003; Thanos, Dimitrakakis, Rice, Gifford, & Volkow, 2005; Vinod, Yalamanchili, et al., 2008) (Table 6). In contrast, CB2R knockout mice showed increased ethanol consumption in two different drinking models (Ortega-Alvaro et al., 2015). Fatty acid amide hydrolase (FAAH) metabolizes endocannabinoids, and FAAH inhibition increases anandamide levels in brain. Increased ethanol intake and preference were reported in FAAH knockout mice on B6 and B6×129/SvJ backgrounds (Basavarajappa, Yalamanchili, Cravatt, Cooper, & Hungund, 2006; Blednov, Cravatt, Boehm, Walker, & Harris, 2007; Vinod, Sanguino, Yalamanchili, Manzanares, & Hungund, 2008).

Table 6. Alcohol consumption in cannabinoid and opioid knockout mice.

Gene Knockout	Background	Operant	2BC	DID	References
Cannabinoid receptor 1 (Cnr1)	B6 × 129/Ola		↓ (6 h)		(Poncelet et al., 2003)
	B6 × 129/Ola		↓ preference		(Lallemand & de Witte, 2005)
	B6		↓ males/females		(Hungund et al., 2003)
	CD1		↓ males/females		(Naassila et al., 2004)
	CD1			↓ (8 h)	(Thanos, Dimitrakakis, et al., 2005)
	B6 and DBA/2J		↓		(Vinod, Yalamanchili, et al., 2008)
	B6		— after 1 week		(Racz et al., 2003)
Cannabinoid receptor 2 (Cnr2)	CD1 and B6	↑ (1 h)	↑		(Ortega-Alvaro et al., 2015)
Fatty acid amide hydrolase (Faah)	B6 × 129/SvJ		↑ males/females		(Blednov et al., 2007)
	B6		↑ females — males		(Basavarajappa et al., 2006)
	B6		↑		(Vinod, Sanguino, et al., 2008)
μ-type opioid receptor (Oprm1)	B6 × 129/Sv	↓ nose-poke ↓ lever operant	↓ post-operant ↓ post-operant ↓ post-forced ethanol		(Roberts et al., 2000)
	B6 × 129/Sv		↓ females — males		(Hall et al., 2001)
	Not specified		↓ after several weeks		(Becker et al., 2002)
	B6			↓ (2 h) — (2h) ADE	(Contet et al., 2014)
	B6			— (4 h)	(van Rijn & Whistler, 2009)
δ-type opioid receptor (Oprd1)	B6			↑ (4 h)	(van Rijn & Whistler, 2009)
	B6Orl × 129/Sv	↑ post-2BC	— naive ↑ post-operant		(Roberts et al., 2001)
κ-type opioid receptor (Oprk1)	B6			↓ (4 h)	(van Rijn & Whistler, 2009)
	B6Orl		↓ males/females		(Kovacs et al., 2005)
β-endorphin (Pomc)	B6		↑ 7%, males/females — 10%, males/females		(Grisel et al., 1999)
	B6		—↑ 2 days, ADE ↑ (2 h) days 1-10 ↑ day 2, ADE		(Grahame, Mosemiller, Low, & Froehlich, 2000)
	B6		↓ females — preference, males		(Racz et al., 2008)
Prodynorphin (Pdyn)	B6		↑ males/females		(Racz, Markert, Mauer, Stoffel-Wagner, & Zimmer, 2013)
	B6		—		(Sperling, Gomes, Sypek, Carey, & McLaughlin, 2010)
	B6		↑		(Femenia & Manzanares, 2012)
	B6 × 129/SvEv-Tac		↓ females — males		(Blednov, Walker, Martinez, & Harris, 2006)
Preproenkephalin (Penk1)	B6		—		(Koenig & Olive, 2002)
	B6		— females ↑ 1-3 weeks, males		(Racz et al., 2008)

–, ↓, ↑: no significant difference, decreased, or increased ethanol intake and/or preference, respectively, in knockout vs. wild-type mice. Male mice were tested unless otherwise indicated. Ethanol intake in the operant and two-bottle choice (2BC) tests was measured in 30-min and 24-h sessions, respectively, unless otherwise indicated. Wild-type, but not knockout mice, showed a stress-induced increase in ethanol intake/preference (Racz et al., 2003; 2008; 2013; Sperling et al., 2010). Drinking in the dark (DID) session times are noted in parenthesis. ADE, alcohol deprivation effect. Recommended mouse protein and gene (in italics) names are from Uniprot. B6 refers to C57BL/6J mice.

Endogenous opioids are expressed throughout the reinforcement circuitry in the brain, and cannabinoid and opioid systems may interact to regulate the effects of drugs of abuse (Charbogne, Kieffer, & Befort, 2014). Table 6 shows the effects of opioid receptor mutations on voluntary alcohol drinking. Mice lacking the mu-opioid receptor (MOR) did not self-administer ethanol (Roberts et al., 2000). This was demonstrated in two different operant procedures (nose-poke and lever press) and in 2BC drinking with or without prior ethanol experience. Other studies using MOR knockout mice reported decreased drinking compared to wild-type mice (Becker et al., 2002; Contet et al., 2014; Hall, Sora, & Uhl, 2001). Decreased consumption was also observed in kappa-opioid receptor (KOR) knockout mice (Kovacs et al., 2005; van Rijn & Whistler, 2009). In contrast, mice lacking delta-type opioid receptors (DORs) showed increased alcohol drinking behavior (Roberts et al., 2001; van Rijn & Whistler, 2009). The effects of deleting β-endorphin, prodynorphin, or preproenkaphalin varied depending on the ethanol concentration, time of exposure, sex, and genotype (Table 6).

Consistent with the preclinical research showing that knockout of select opioid receptors reduces drinking in mice, human genetic studies further support a role for polymorphisms in genes coding for MOR1, DOR1, KOR1, and other opioid receptors in clinical populations of alcoholics (Levran, Yuferov, & Kreek, 2012). In addition, opioids are used in the treatment of AUD. The pan-opioid antagonist naltrexone is already FDA approved to treat AUD and may reduce craving and relapse to heavy drinking in humans. Nalmefene, another pan-opioid receptor antagonist structurally similar to naltrexone, is currently used to treat opioid overdose in the U.S. and is approved in Europe for as-needed use to reduce heavy drinking. It has several advantages over naltrexone, including longer duration of action, lack of dose-dependent liver toxicity, and higher affinity of binding to MORs and KORs. Nalmefene reduced the total amount of alcohol consumed and the number of heavy drinking days in alcohol-dependent patients (van den Brink et al., 2014).

Immune-related Genes

The interplay between brain, behavior, and immune responses in the etiology and progression of drug abuse is a current area of interest in addiction research (http://www.arcr.niaaa.nih.gov/arcr/arcr372/toc37_2.htm). The neuroimmune system, encompassing innate immune signaling within the peripheral and central nervous systems, is important in the pathophysiology and potential treatment of alcohol abuse and dependence (Crews & Vetreno, 2015; Mayfield, Ferguson, & Harris, 2013; Robinson et al., 2014). Alcohol may increase neuroimmune gene and protein expression through both peripheral-central signaling molecules and direct actions in the brain (Crews & Ventreno, 2015).

Deletion of many different immune-related genes lowered ethanol intake and preference in male and female mice on a B6 background (Table 7). In some cases, the effect depended on the drinking test used. Deletion of certain chemokine and chemokine receptor genes reduced ethanol preference and intake (Blednov et al., 2005). Knockout mice were also used to study other immune-associated genes previously linked to alcohol consumption in a gene expression analysis of mouse brain (Mulligan et al., 2006). The following null mutations were examined based on this transcriptome meta-analysis: B2m (beta-2 microglobulin), Cd14 (cluster of differentiation 14), Il1rn (interleukin 1 receptor antagonist), Il6 (interleukin 6), Ctss (cathepsin S), and Ctsf (cathepsin F) (Blednov et al., 2012). These candidate targets were subsequently validated in behavioral tests because knockout of these individual genes reduced ethanol consumption and preference in 2BC tests (Blednov et al., 2012). In contrast, a transgenic mouse line overexpressing Il6 showed increased alcohol preference in females but not males (Harris & Blednov, 2013). C3H/HeJ mice are naturally TLR4-deficient and showed decreased operant self-administration of ethanol compared to the control strain (Harris & Blednov, 2013); however, ethanol consumption was not changed in TLR4 knockout mice in a 2BC test (Pascual, Balino, Alfonso-Loeches, Aragon, & Guerri, 2011).

Table 7. Alcohol consumption in immune-related receptor knockout mice.

Gene Knockout	Background	2BC	DID	References
Cathepsin F (Ctsf)	B6 × 129/SvJ	↓malesn — females		(Blednov et al., 2012)
Cathepsin S (Ctss)	B6	↓males/females	— 2BC, males ↓ 1B, males ↓ 1B, 2BC, females	(Blednov et al., 2012)
C-C motif chemokine 2 (Cc12)	B6	— males ↓ females		(Blednov et al., 2005)
C-C motif chemokine 3 (Cc13)	B6	↓ males/females		(Blednov et al., 2005)
C-C chemokine receptor type 2 (Ccr2)	B6	↓ males/females		(Blednov et al., 2005)
C-C chemokine receptor type 5 (Ccr5)	B6	— males/females		(Blednov et al., 2005)
Apolipoprotein E (Apoe)	B6	— females		(Bechtholt, Smith, Raber, & Cunningham, 2004)
Monocyte differentiation antigen CD14 (Cd14)	B6	↓ males/females	— 1B, 2BC, males — 1B, females ↓ 2BC, females	(Blednov et al., 2012)
Interleukin 1 receptor antagonist (Il1rn)	B6 × 129/SvJ	↓ males/females	↓ 1B, 2BC, males ↓ 1B, 2BC, females	(Blednov et al., 2012)
Interleukin-6 (Il6)	B6	↓ males/females	— 1B, 2BC, males — 1B, 2BC, females	(Blednov et al., 2012)
β-2-microglobulin (B2m)	B6	↓ males/females	— 1B, 2BC, males — 1B, 2BC, females	(Blednov et al., 2012)
B2 bradykinin receptor (Bdkrb2)	B6	—		(Maul et al., 2005)
Toll-like receptor 4 (Tlr4)	B6	—		(Pascual et al., 2011)

–, ↓, ↑: no significant difference, decreased ethanol intake and/or preference, or increased ethanol intake and/or preference, respectively, in knockout vs. wild-type mice. Male mice were tested unless indicated otherwise. Ethanol intake in the two-bottle choice (2BC) tests was measured in 24-h sessions. Drinking in the dark (DID) session times were 2 and 4 h; 1B, one bottle. Recommended mouse protein and gene (in italics) names are from Uniprot. B6 refers to C57BL/6J mice.

Pharmacological inhibition of immune signaling reduces alcohol drinking in mice, whereas activation of the innate immune system using lipopolysaccharide produces prolonged increases in drinking (Blednov, Benavidez, et al., 2011). Immune molecules, acting peripherally and/or centrally, appear to be involved in preference for alcohol and the behavioral evidence agrees with expression profiling of mouse brain (Robinson et al., 2014). Unique gene expression patterns in human post-mortem brain provide additional support for immune gene and related network effects in alcohol-dependent individuals (Liu et al., 2006; Ponomarev, Wang, Zhang, Harris, & Mayfield, 2012). Furthermore, genetic association studies in humans link immune-associated genes with alcohol dependence (Blednov, Benavidez, Black, Ferguson, et al., 2015). The use of agents that regulate immune responses as potential therapeutics for AUD is discussed in the Medication Development section later in the chapter.

Ion Channels

Some of the rapid-onset actions of alcohol are likely mediated by direct action on ion channels (Howard, Trudell, & Harris, 2014; Trudell et al., 2014). In addition to the prominent and well-studied neurotransmitter systems in alcohol dependence discussed previously (e.g., GABA, glutamate, dopamine, 5-HT), other ion channels that have been implicated in alcohol intake and preference are described below.

Transient receptor potential (TRP) cation channels

TRP channels are a group of non-selective ion channels that mediate sensations, tastes, temperature, pain, and other chemical and physical stimuli. There is a positive correlation between ethanol intake and consumption of sweet solutions in rodents and humans (Kampov-Polevoy et al., 2014). Deletion of TRP cation channel subfamily M, member 5 substantially decreased ethanol intake and preference (Blednov et al., 2008) (Table 8). Altered taste perception may explain the reduced preference for both saccharin and ethanol. Mice lacking TRP cation channel subfamily V, member 1 (TRPV1), also known as the vanilloid or capsaicin receptor, showed increased ethanol preference but no change in saccharin or quinine preference (Blednov & Harris, 2009). The endocannabinoid, anandamide, is an endogenous activator of this channel, and ethanol also directly activates TRPV1 at high concentrations. The increased ethanol consumption in Trpv1 knockout mice may simply be due to a reduction in the ethanol-induced burning sensation, although some central actions, including endocannabinoid signaling, may be involved.

Table 8. Alcohol consumption in ion channel mutant mice.

Gene Knockout/Overexpression*	Background	2BC	CIE	DID	References
Transient receptor potential cation channel, subfamily M member 5 (Trpm5)	B6	↓			(Blednov et al., 2008)
Transient receptor potential cation channel subfamily V member 1 (Trpv1)	B6 × 129X1/SvJ	↑			(Blednov & Harris, 2009)
G protein-activated inward rectifier potassium channel 2, GIRK2 (Kcnj6)	B6 × 129Sv	— withbottle alternation; males/females ↑ without bottle alternation; females			(Blednov et al., 2001)
G protein-activated inward rectifier potassium channel 3, GIRK3 (Kcnj9)	B6	↑ (2 h) —			(Herman et al., 2015)
Voltage-dependent N-type calcium channel subunit α-1B (Cacna1b)	B6 × 129SvJae	↓			(Newton et al., 2004)
Calcium-activated potassium channel subunit β-1, BK β1 (Kcnmb1)	B6	— continuous, intermittent	↑a	— (2 h)	(Kreifeldt et al., 2013)
Calcium-activated potassium channel subunit β-4, BK β4 (Kcnmb4)	B6	— continuous, intermittent	↓a	— (2 h)	(Kreifeldt et al., 2013)
	B6			↑ (2 h) males/females	(Martin et al., 2008)
P2X purinoceptor 4 (P2rx4)	B6	↑ intake		↑ (4 h) 5%, 20% — (4 h) 10%	(Wyatt et al., 2014)
Glycine receptor subunit α-2 (Glra2)	B6	↓ — intermittent		— (2, 4 h) 1B	(Blednov, Benavidez, Black, Leiter, et al., 2015)
Glycine receptor subunit α-3 (Glra3)	B6	— ↑ intermittent		— (2, 4 h) 1B	(Blednov, Benavidez, Black, Leiter, et al., 2015)
Nicotinic acetylcholine receptor (nAChR) α4 subunit (Chrna4)	B6			↓ (2 h) 20% — (2 h) 2%	(Hendrickson et al., 2010)
nAChR α5 subunit (Chrna5)	B6			— (4 h)	(Santos et al., 2013)
nAChR α6 subunit (Chrna6)	B6	— males/females			(Kamens et al., 2012)
nAChR α7 subunit (Chrna7)	B6	↓ females — males			(Kamens et al., 2010)
nAChR β2 subunit (Chrnb2)	B6	— males/females			(Kamens et al., 2010)
	B6	— intermittent			(Dawson et al., 2013)
nAChR β3 subunit (Chrnb3)	B6	— males/females			(Kamens et al., 2012)
nAChR α5α3β4*	B6SJLF1/J	↓ intake			(Gallego et al., 2012)

^*indicates gene overexpression;

–, ↓, ↑: no difference, decreased, or increased ethanol intake and/or preference, respectively, in mutant vs. wild-type mice. Male mice were tested unless otherwise indicated. Ethanol intake in the two-bottle (2BC) tests was measured in 24-h sessions unless otherwise indicated. Drinking session times for the other tests are indicated in parenthesis. CIE, chronic intermittent ethanol; DID, drinking in the dark; 1B, one bottle.

^aThese tests involved weekly limited access (2 h/day) 2BC drinking alternated with weekly CIE vapor (16 h/day) to create alcohol dependence/withdrawal. Deletion of BK β4 attenuated, while deletion of BK β1 accelerated, the escalation of ethanol drinking in dependent mice during withdrawal from CIE. Recommended mouse protein and gene (in italics) names are from Uniprot. B6 refers to C57BL/6J mice.

N-type calcium channels

Voltage-dependent calcium channels regulate neuronal excitability, neurotransmitter release, and gene expression. Several types exist with unique pharmacological and electrophysiological characteristics. L-, T-, and N-type voltage-gated calcium channels have all been implicated in the behavioral effects of ethanol. Reduced ethanol intake and preference were reported in mice lacking functional α_1B subunits of N-type calcium channels (Newton et al., 2004) (Table 8), suggesting these subunits as a potential target to control drinking. The calcium channel blocker gabapentin targets channels with α₂δ subunits (Geisler, Schopf, & Obermair, 2015) and, as previously mentioned, offers some promise as a therapeutic for AUD (Mason et al., 2014).

GIRK channels

G protein-coupled inwardly rectifying K⁺ (GIRK) channels are widely expressed in brain and are activated by ligand-stimulated G protein-coupled receptors, resulting in an outward K⁺ current that hyperpolarizes neuronal membranes and decreases neuronal excitability. Genetic evidence from mouse models has provided extensive insight into the significance of GIRK channels in drug addiction, and their role in responses to ethanol and other drugs of abuse is reviewed in (Bodhinathan & Slesinger, 2014) and (Mayfield, Blednov, & Harris, 2015).

As shown in Table 8, ethanol consumption and preference did not differ in wild-type and Girk2 knockout mice in the standard 2BC test, where the bottle positions were alternated daily to control for position preferences. However, when the ethanol bottles were always available in the preferred location, Girk2 knockout mice consumed more ethanol compared to wild-type (Blednov, Stoffel, Chang, & Harris, 2001). Moreover, single nucleotide polymorphisms of the KCJN6 gene (GIRK2) have been associated with alcohol dependence in humans (Clarke et al., 2011; Kang et al., 2012). Deletion of Girk3 increased limited- but not continuous-access voluntary drinking, and overexpression of Girk3 in the VTA reversed the binge-drinking phenotype and reduced drinking in wild-type mice (Herman et al., 2015). These results point to a role for GIRK subunits in the rewarding properties of ethanol and as a potential target for regulating binge-like drinking.

BK channels

BK channels are calcium-activated potassium channels that are characterized by their large conductance of potassium ions. The β subunits influence BK channel responses to acute and chronic ethanol, and ethanol action on one system cannot easily be extrapolated to another. In 2BC continuous and intermittent access tests, non-dependent mice lacking BK β1 or β4 subunits did not alter their drinking compared to wild-type (Kreifeldt, Le, Treistman, Koob, & Contet, 2013) (Table 8). Weeks of voluntary 2BC drinking alternated with weeks of CIE vapor escalated drinking in dependent wild-type mice to a greater extent than BK β4 knockout mice. In contrast, β1 knockout mice drank more after fewer CIE cycles than did wild-type mice. Thus loss of β4 decreased, while loss of β1 increased, the escalation of drinking during withdrawal from CIE (Kreifeldt et al., 2013).

Nicotinic acetylcholine receptors

Neuronal nicotininc acetylcholine receptors (nAChRs) are involved in several alcohol-mediated behaviors (Rahman, Engleman, & Bell, 2014). Varenicline (which has partial agonist activity at nAChRs and is used in smoking cessation programs) reduced drinking in both preclinical and clinical studies, and human genetic studies showed a strong association of polymorphisms in CHRNA5 (which encodes the α5 subunit) with risk of developing alcohol dependence, reviewed in (Rahman et al., 2014). Mice lacking Chrna5 did not differ from wild-type in ethanol consumption (Santos, Chatterjee, Henry, Holgate, & Bartlett, 2013), but transgenic mice overexpressing α5α3β4 showed reduced intake (Gallego et al., 2012) (Table 8). Knockout of α6 nACh subunits had no effect on 2BC drinking (Kamens, Hoft, Cox, Miyamoto, & Ehringer, 2012). However, deletion of α4 decreased DID of a high ethanol concentration (Hendrickson, Zhao-Shea, Pang, Gardner, & Tapper, 2010), and loss of α7 decreased 2BC drinking in female mice (Kamens, Andersen, & Picciotto, 2010). Knockout of β2 (Dawson, Miles, & Damaj, 2013; Kamens et al., 2010) or β3 (Kamens et al., 2012) nAChR subunits had no effect in 2BC tests in male or female mice.

Glycine receptors

Supraspinal glycine receptors are found in reward pathways and play a key role in alcohol inhibitory responses in some neurons (Perkins, Trudell, Crawford, Alkana, & Davies, 2010). The glycine receptor α1 subunit has proven difficult to engineer and studies are complicated by the reduced glycinergic function and lethality of homozygous mutants. Deletion of glycine receptor subunits α2 or α3 had no effect on limited-access DID, while 2BC continuous-access drinking decreased in α2 but not α3 null mutants (Blednov, Benavidez, Black, Leiter, et al., 2015) (Table 8). However, intermittent-access drinking increased in α3 but not α2 null mutants. A linkage analysis in humans also supports a role for the α3 subunit in alcohol dependence (Han, Gelernter, Kranzler, & Yang, 2013).

Protein Kinases

In addition to direct effects on ion channels, ethanol indirectly modulates channel function via phosphorylation and other post-translational processing mechanisms (Trudell et al., 2014). Mice lacking protein kinase C type ε (PKCε) drank less ethanol than wild-type, and this effect has been observed in different drinking tests across multiple labs (Besheer, Lepoutre, Mole, & Hodge, 2006; Choi, Wang, Dadgar, Chang, & Messing, 2002; Hodge et al., 1999; Olive et al., 2005; Olive, Mehmert, Messing, & Hodge, 2000; Wallace et al., 2007) (Table 9). Selective knockdown of PKCε in the amygdala also reduced 2-hour ethanol consumption in a DID model (Lesscher et al., 2009). PKCε is known to reduce the positive allosteric effects of benzodiazepines and ethanol at GABA_A receptors, which may in turn regulate reward signaling. The effects of deleting other PKC types on alcohol intake and preference have not been studied as extensively, although drinking in 2BC tests increased in mice lacking PKCγ (Bowers & Wehner, 2001) or ζ (A. M. Lee et al., 2014).

Table 9. Alcohol consumption in kinase mutant mice.

Gene Knockout/Overexpression*	Background	Operant	2BC	DID	References
cAMP-dependent protein kinase type II-β regulatory subunit (Prkar2b)	B6 × 129/SvJ		↑ males/females		(Thiele, Willis, et al., 2000)
	B6	genotype/sex interactions	↑ males/females		(Ferraro, Sparta, Knapp, Breese, & Thiele, 2006)
	B6		↑ males/females		(Fee et al., 2004)
	B6 × 129/SvEv		↑ males/females		(Fee et al., 2004)
Protein kinase C (PKC) ε type (Prkce)	B6 × 129/SvJae	↓ ↓ ADE			(Olive et al., 2000)
	B6 × 129/SvJae× FVB/N		↓		(Choi et al., 2002)
	B6 × 129/SvJae		↓ (23 h)		(Hodge et al., 1999)
	B6 × 129/SvJae		↓ (16 h)		(Olive et al., 2005)
	B6 × 129S4		↓ ↓ post-tolerance		(Wallace et al., 2007)
	B6 ×129/SvJae		↓		(Besheer et al., 2006)
	B6 × 129S4/SvJae			↓ (2 h) amygdala	(Lesscher et al., 2009)
PKC γ type (Prkcg)	B6 × 129/SvEvTac		↑ males/females		(Bowers & Wehner, 2001)
PKC ζ type (Prkcz)	B6 × 129S6/SvEvTac		— continuous ↑ (4, 24 h) intermittent	— (4 h)	(A. M. Lee et al., 2014)
Tyrosine-protein kinase Fyn (Fyn)	B6 × 129SF2/J		↓		(Boehm, Peden, Chang, Harris, & Blednov, 2003)
	129/SvJ		— (23 h)		(Yaka, Tang, Camarini, Janak, & Ron, 2003)
	B6 × CBA		— — ADE — post-stress		(Cowen, Schumann, Yagi, & Spanagel, 2003)
Fyn*	B6 × 129/Sv		↓ males/females		(Boehm, Peden, et al., 2004)
	B6		↓		(Stork, Kojima, Stork, Kume, & Obata, 2002)
cGMP-dependent protein kinase 2 (Prkg2)	B6N and 129/SvN		↑		(C. Werner et al., 2004)
Serine/threonine-protein kinase, TAO2 (Taok2)	B6			↑ (4 h)	(Kapfhamer et al., 2013)
ALK tyrosine kinase receptor (Alk)	B6			↑ (4 h) intermittent	(Lasek et al., 2011)

^*indicates gene overexpression;

–, ↓, ↑: no significant difference, decreased, or increased ethanol intake and/or preference, respectively, in mutant vs. wild-type mice. Male mice were tested unless indicated otherwise. Ethanol intake in the two-bottle choice (2BC) tests was measured in 24-h sessions, unless otherwise indicated. Session times for operant tests were 16 h, and drinking in the dark (DID) session times are indicated in parenthesis. ADE, alcohol deprivation effect. Recommended mouse protein and gene (in italics) names are from Uniprot. B6 refers to C57BL/6J mice.

Enzymes

Table 10 shows the effects of other assorted enzymes on ethanol drinking in mice. In particular, aldehyde dehydrogenase (ALDH) is one of the few known genes to affect risk of developing AUD in humans (C. H. Chen, Ferreira, Gross, & Mochly-Rosen, 2014). ALDH2 plays a major role in the detoxification of ethanol-derived acetaldehyde, and inhibition of ALDH is the mechanism of action of disulfiram, an FDA-approved drug for AUD. A mutation in ALDH2 produces an enzyme incapable of metabolizing ethanol, resulting in severe adverse reactions to even small amounts of ingested alcohol in individuals who are homozygous for the mutation. UChA and UChB rat lines show similar Aldh2 polymorphisms that affect alcohol consumption (Quintanilla, Israel, Sapag, & Tampier, 2006). Knockout of the Aldh2 gene in mice was also shown to limit drinking in 2BC tests (Fernandez et al., 2006; Isse et al., 2002). Thus, genetic polymorphisms that alter ethanol metabolism and susceptibility to its effects may decrease ethanol-induced metabolic tolerance and risk for developing AUD.

Table 10. Alcohol consumption in enzyme knockout mice.

Gene Knockout	Background	Operant	2BC	DID	References
Adenylate cyclase (AC) type 1 (Adcy1)	B6		—		(Maas et al., 2005)
AC type 5 (Adcy5)	B6		↑ males/females		(Kim, Kim, Baek, Lee, & Han, 2011)
AC type 8 (Adcy8)	B6		↓		(Maas et al., 2005)
Adcy1/Adcy8 double knockout	B6		↓		(Maas et al., 2005)
Pituitary adenylate cyclase-activating polypeptide (Adcyap1)	Crlj:CD1		↑		(Tanaka et al., 2010)
Aldehyde dehydrogenase, mitochondrial (Aldh2)	B6		↓		(Isse et al., 2002)
	B6 × 129Sv/lex		↓ sex not specified		(Fernandez et al., 2006)
Glutamyl aminopeptidase (Aminopeptidase A) (Enpep)	B6 × 129Sv		— — post-stressa		(Faber et al., 2006)
Catechol-O-methyltransferase (Comt)	B6		↑ males — females		(Tammimaki, Forsberg, Karayiorgou, Gogos, & Mannisto, 2008)
Amine oxidase [flavin-containing] A (Maoa)	Tg8 and C3H/HeJ		— (2, 24 h)		(Popova, Vishnivetskaya, Ivanova, Skrinskaya, & Seif, 2000)
Neprilysin, NEP (Mme)	B6N		— ↑ post-stress		(Maul et al., 2012)
	Not specified		↑		(Siems et al., 2000)
Nitric oxide synthase, brain (Nos1)	B6 × 129X1/SvJ		↑ 8-16%, sex not specified		(Spanagel et al., 2002)
Cytochrome P450 2E1 (Cyp2e1)	129S1/SV-Ter		↓ preference 4-8%, females		(Correa et al., 2009)
Protein phosphatase 1 regulatory subunit 1B, DARPP-32 (Ppp1r1b)	B6	↓ (23 h)			(Risinger, Freeman, Greengard, & Fienberg, 2001)
Histidine decarboxylase (Hdc)	129/SvB6		—	— (4 h), males/females	(Vanhanen et al., 2013)
Tyrosine-protein phosphatase non-receptor type 5, STEP (Ptpn5)	B6		↑		(Legastelois, Darcq, Wegner, Lombroso, & Ron, 2015)
Ubiquitin carboxyl-terminal hydrolase 46 (Usp46)	B6		↓		(Imai, Kano, Nonoyama, & Ebihara, 2013)

–, ↓, ↑: no significant difference, decreased, or increased ethanol intake and/or preference, respectively, in knockout vs. wild-type mice. Male mice were tested unless otherwise indicated.

^aSocial stress reduced alcohol consumption in both knockout and wild-type mice, but there was no genotype difference. Ethanol intake in the two-bottle choice (2BC) tests was measured in 24-h sessions, unless indicated otherwise. Drinking session times for the other tests are indicated in parenthesis. DID, drinking in the dark; 1B, one bottle. Recommended mouse protein and gene (in italics) names are from Uniprot. B6 refers to C57BL/6J mice.

Neuropeptides/Hormones

Table 11 summarizes the effects of deletion or overexpression of classical or putative neuropeptides and hormones and their receptors on voluntary ethanol administration in mice. Of the genes represented here, the corticotropin releasing factor/urocortin family and other stress-related neuromodulators are promising for future studies of genetic determinants of AUD (Schank, Ryabinin, Giardino, Ciccocioppo, & Heilig, 2012).

Table 11. Alcohol consumption in neuropeptide/hormone mutant mice.

Gene Knockout/Overexpression*	Background		Operant	2BC	DID	References
Neuropeptide Y, NPY (Npy)	129/SvEv			— 3-10%, ↑ 20%		(Thiele, Miura, Marsh, Bernstein, & Palmiter, 2000)
	B6 × 129Sv			↑
Npy*	B6 × 129Sv			↓		(Thiele, Marsh, Ste Marie, Bernstein, & Palmiter, 1998)
NPY receptor type 1 (Npy1r)	B6			↑ 3-10%, males ↑ 10%, females		(Thiele, Koh, & Pedrazzini, 2002)
NPY receptor type 2 (Npy2r)	129/SvJ × Balb/cJ			↓		(Thiele, Naveilhan, & Ernfors, 2004)
	Balb/cJ			—		(Thiele et al., 2004)
NPY receptor Y5 (Npy5r)	129/SvEv			—		(Thiele, Miura, et al., 2000)
Agouti-related protein (Agrp)	B6		↓ (2 h), males/females	↓ (2 h), females	— (2 h), males/females ↓ (4 h), males/females	(Navarro, Cubero, Ko, & Thiele, 2009)
Angiotensinogen (Agt)	Not specified			↓		(Maul et al., 2001)
Agt*	Not specified			↑		(Maul et al., 2001)
Type-1A angiotensin II receptor, AT1A (Agtr1a)	Not specified			↓ sex not specified		(Maul et al., 2005)
Agtr1a*	Not specified			—		(Moore, Krstew, Kirchhoff, Davisson, & Lawrence, 2007)
Type-2 angiotensin II receptor, AT2 (Agtr2)	Not specified			— sex not specified		(Maul et al., 2005)
Cholecystokinin receptor type A, CCK-A (Cckar)	B6			— preference		(Miyasaka et al., 2005)
Gastrin/cholecystokinin type B receptor, CCK-B (Cckbr)	B6			—		(Miyasaka et al., 2005)
	B6			— males — preference, females		(Abramov et al., 2006)
Cocaine- and amphetamine-regulated transcript protein (Cartpt)	B6			↓, males/females		(Salinas, Nguyen, Ahmadi-Tehrani, & Morrisett, 2014)
Corticotropin-releasing factor (Crh)	B6 × 129S			↑ (2, 23 h)		(Olive et al., 2003)
	B6				↓ (2, 4 h), males/females	(Kaur, Li, Stenzel-Poore, & Ryabinin, 2012)
Corticotropin-releasing factor (CRF) receptor 1 (Crhr1)	B6				↓ (2, 4 h), males/females	(Kaur et al., 2012)
	B6 × 129SvJ		— (1 h) ↓ (1h) ADE			(Chu, Koob, Cole, Zorrilla, & Roberts, 2007)
	129/Ola × CD1			— ↑ post-stress		(Sillaber et al., 2002)
	B6, males/females			— 3-10%, ↓ 20% — (21 h) 10%, pre-/post-stress ↓ (21 h) 10%, chronic stress		(Pastor et al., 2011)
	B6				↓ intake (2, 4 h), males/females	(Giardino & Ryabinin, 2013)
	129 × 1/SvJ × CD1			— ↑ post-stress ↑ escalation post-dependence — ADE		(Molander et al., 2012)
Crhr1NestinCre	129S2/Sv × B6 × (B6 × SJL)			— ↓ post-stress ↓ escalation post-dependence — ADE		(Molander et al., 2012)
CRF receptor 2 (Crhr2)	B6				↓ (2 h) day 1, males/females	(Kaur et al., 2012)
	B6			— females	↑ (30 min), males/females — (2 h), males/females	(Sharpe et al., 2005)
Crhr1/Crhr2 double knockout	B6, males/females			— (21 h) 10%, pre-/post-stress ↓ (21 h) 10%, chronic stress		(Pastor et al., 2011)
	B6				— (2, 4 h), males/females	(Kaur et al., 2012)
Galanin peptides (Gal)	129Ola/Hsd			↓ 15%, females — males		(Karatayev et al., 2010)
Gal*	B6			↑ 15%, pre-/post-food deprivation, males; — females		(Karatayev, Baylan, & Leibowitz, 2009)
Neurokinin-1 receptor, Substance P receptor (Tacr1)	B6			↓		(Thorsell, Schank, Singley, Hunt, & Heilig, 2010)
Nociceptin receptor (Oprl1)	B6			↓ 20%, females		(Sakoori & Murphy, 2008)
Relaxin-3 receptor 1 (Rxfp3)	B6			— ↓ post-stress		(Walker et al., 2015)
Melanin-concentrating hormone receptor 1 (Mchr1)	B6			↑ males — females	— (1 h), males/females	(Duncan et al., 2007)
Neurotensin receptor type 1 (Ntsr1)	B6 × 129X1/SvJ			↑		(M. R. Lee et al., 2010)
Neurotensin receptor type 2 (Ntsr2)	B6 × 129X1/SvJ			↑		(M. R. Lee, Hinton, Unal, Richelson, & Choi, 2011)
Urocortin (Ucn)	B6				— (2, 4 h), males/females	(Kaur et al., 2012)
	B6			↓ preference		(Giardino, Cocking, Kaur, Cunningham, & Ryabinin, 2011)
Vasopressin V1a receptor (Avpr1a)	B6Cr Slc × 129/Sv and B6Cr Slc			↑ males> females		(Sanbe et al., 2008)
	B6 × 129/Sv-CP, females			— homozygous mutant — post-stress		(Caldwell et al., 2006)
Vasopressin V1b receptor (Avpr1b)	B6 × 129/SvJ, females			— — post-stress		(Caldwell et al., 2006)
Adiponectin receptor protein 2 (Adipor2)	B6				— (2 h) ↓ (2 h) CIE	(Repunte-Canonigo et al., 2010)
Atrial natriuretic peptide receptor 1 (Npr1)	B6 × 129/SvJ			— ↑ post-stress		(Mutschler et al., 2010)
Ghrelin (Ghrl)	B6				— (90 min)	(Jerlhag, Landgren, Egecioglu, Dickson, & Engel, 2011)
	Not specified			↓		(Bahi et al., 2013)
Leptin (Lep)	B6-Lepob			↓ males/females		(Blednov, Walker, & Harris, 2004)
Leptin receptor (Lepr)	B6-m Leprdb/J			↓ males/females		(Blednov, Walker, & Harris, 2004)
Melanocortin receptor 3 (Mc3r)	B6			— males/females		(Navarro et al., 2005)
	B6				↓ (1 h) — (4 h), males/females	(Olney, Sprow, Navarro, & Thiele, 2014)
Melanocortin receptor 4 (Mc4r)	B6			— (6, 24 h), males/females		(Navarro et al., 2011)

^*indicatesgene overexpression;

–, ↓, ↑: no significant difference, decreased, or increased intake and/or preference, respectively, in mutant vs. wild-type/control mice. Male mice were tested unless indicated otherwise. Ethanol intake in the two-bottle choice (2BC) tests was measured in 24-h sessions unless indicated otherwise. Drinking session times for operant and drinking in the dark (DID) tests are indicated in parenthesis. Stress and alcohol dependence tend to increase ethanol intake in both control and knockout mice, and the arrows indicate if the drinking in knockout mice was lower or higher compared to control under these conditions (Molander et al., 2012); however, the knockout mice in Pastor et al. (2011) showed reduced intake following chronic stress, whereas wild-type mice showed increased intake. ADE, alcohol deprivation effect; CIE, chronic intermittent ethanol vapor. Recommended mouse protein and gene (in italics) names are from Uniprot. B6 refers to C57BL/6J mice. Classical neuropeptides are listed followed by a putative neuropeptide and hormones and hormone receptors.

Other Gene Targets

Table 12 represents an assortment of genes that did not specifically fit into the previous categories. These genes are associated with synaptic function, development, circadian regulation, and other cellular regulatory functions but are not discussed individually herein. One of the most pronounced phenotypes among mutant mice is the almost complete blockade of alcohol consumption in mice lacking any one of three taste genes (Gnat3, Tas1r3, Trpm5) (Blednov et al., 2008).

Table 12. Alcohol consumption in other mutant mice.

Gene Knockout/Overexpression*	Background	Operant	2BC	DID	References
α-synuclein (Snca)	B6		↑		(Lopez-Jimenez et al., 2013)
β-arrestin-1 (Arrb1)	B6		—		(Bjork et al., 2008)
β-arrestin-2 (Arrb2)	B6		↓ 9-15%		(Bjork et al., 2008)
	B6		↑		(H. Li, Tao, Ma, Liu, & Ma, 2013)
Gap junction delta-2 protein, Connexin-36 (Gjd2)	B6			↓ (2 h) 20%	(Steffensen et al., 2011)
G1/S-specific cyclin-D2 (Ccnd2)	129X1/SvJ × B6 × BALB/cAnNCrl		— 2%, 4% ↑ 8%, 16%		(Jaholkowski et al., 2011)
α-gustducin (Gnat3)	B6		↓		(Blednov et al., 2008)
Histamine H3 receptor (Hrh3)	B6		↓	↓ (2, 4 h)	(Nuutinen et al., 2011)
LIM domain only protein 3 (Lmo3)	B6		— males/females	↑ (2, 4 h), males/females	(Savarese, Zou, Kharazia, Maiya, & Lasek, 2014)
Metallothionein-1/2 (Mt1, Mt2 double knockout)	129S7/SvEvBrd and 129S1/SvImJ		↑ males/females		(Loney, Uddin, & Singh, 2006)
Period circadian protein homolog 1 (Per1)	Per1Brdm1 B6-Tyrc-Brd × 129S7		— drink-o-meter ↑ stress		(Dong et al., 2011)
	B6-Tyrc-Brd × 129SvEvBrd	—	— — ADE		(Zghoul et al., 2007)
Period circadian protein homolog 2 (Per2)	Per2Brdm1 B6-Tyrc-Brd × 129SvEvBrd	↑	↑ 8-16% ↓ after acamprosate		(Spanagel et al., 2005)
	B6		↑ drink-o-meter		(A. J. Brager et al., 2011)
	B6		↑		(A. Brager et al., 2011)
Disks large homolog 4, PSD-95 (Dlg4)	B6		↓ males/females ↑ day 1, ADE		(Camp et al., 2011)
Protein fosB (Fosb)	129Sv × BALB/c		— males/females		(Korkosz et al., 2004)
Protransforming growth factor α (Tgfa)*	CD1		↑		(Hilakivi-Clarke & Goldberg, 1995)
Epidermal growth factor receptor kinase substrate 8 (Eps8)	B6		↑ males/females		(Offenhauser et al., 2006)
Ras-related protein Rab-3A (Rab3a)	B6		— (23 h)		(Kapfhamer et al., 2008)
Ras-specific guanine nucleotide-releasing factor 2 (Rasgrf2)	B6		↓		(Stacey et al., 2012)
Regulator of G-protein signaling 6 (Rgs6)	B6 × 129/Sv		↓ males/females		(Stewart et al., 2015)
Taste receptor type 1 member 3 (Tas1r3)	B6		↓		(Blednov et al., 2008)
	B6		↓ males/females		(Brasser, Norman, & Lemon, 2010)
Trace amine-associated receptor 1 (Taar1)	B6 × 129S1/Sv		↑ females		(Lynch et al., 2013)
Protein unc-79 homolog (Unc79)	B6 and B6 × DBA/2J		↑		(Speca et al., 2010)

^*indicates gene overexpression;

–, ↓, ↑: no significant difference, decreased, or increased ethanol intake and/or preference, respectively, in mutant vs. wild-type mice. Male mice were tested unless indicated otherwise. Ethanol intake in the operant and two-bottle choice (2BC) tests was measured in 30-min and 24-h sessions, respectively, unless otherwise indicated. Drinking in the dark (DID) session times are indicated in parenthesis. ADE, alcohol deprivation effect. Recommended mouse protein and gene (in italics) names are from Uniprot. B6 refers to C57BL/6J mice.

Concluding Remarks

Mouse models of voluntary ethanol administration have been instrumental for profiling putative behavioral and genetic determinants in human alcoholics, who exhibit excessive consumption as a hallmark of the disease. The impact of more than 150 genes on alcohol consumption has been evaluated by construction of mutant mice. The global knockout strategy has been used extensively in addiction research to link proteins with behavior, and most studies presented in this chapter used this approach. Collectively speaking, the studies suggest the potential involvement of a large number of different mouse genes in voluntary alcohol drinking. However, many of the effects are modest and global knockout or overexpression of a specific gene does not determine if the effect on drinking is directly linked to effects of decreased or increased expression of the respective protein. Other technologies such as knockin animal models and conditional and cell-specific knockouts will provide more discriminating tools for linking ethanol-sensitive sites on proteins to behaviors (Blednov, Borghese, et al., 2011). The goal of the knockin approach is to remove ethanol action on a specific protein without otherwise altering its function, but this strategy (like the knockout model) can still be complicated by compensatory actions from other proteins. Selectively targeting an ethanol site could also affect other modulatory sites of the protein despite the appearance of normal function (Harris, Osterndorff-Kahanek, Ponomarev, Homanics, & Blednov, 2011).

When focusing on genes that have more profound effects on consumption and have consistently been implicated across multiple drinking tests, the number of targets diminishes. For example, null mutant mice for specific GABA_A (Table 1), dopamine (Table 3), cannabinoid and opioid receptors (Table 6), immune-related genes (Table 7), and PKCε (Table 9) have demonstrated decreased drinking and preference in several drinking tests. These systems either have established roles in addiction research or represent emerging targets, as in the case of immune pathways. Further consideration should be given to the impact of the drinking paradigms on peripheral and central genes. For example, even when similar amounts of total ethanol were consumed in the continuous and intermittent access tests by B6 mice, brain and liver transcriptomes were differentially affected (Osterndorff-Kahanek, Ponomarev, Blednov, & Harris, 2013). The distinct genomic effects induced by the alcohol exposure protocol may explain the differing effects observed in some drinking tests. The peripheral effects of alcohol may be particularly important for neuroimmune modulation of drinking because production of cytokines outside the brain can influence brain function. This peripheral-central neuroimmune communication may also be important in alcohol craving and dependence in humans (Leclercq, De Saeger, Delzenne, de Timary, & Starkel, 2014).

Medication Development

FDA-approved drugs for AUD have provided only modest benefit and are not routinely prescribed, and so the search continues for more effective drugs. Identifying existing drugs that could be repurposed to treat AUD is a current goal for researchers and, if successful, would fast track therapeutic options for the disease. A strategy for prioritizing relevant genes from a large list of potential targets is to examine the preclinical evidence in combination with genetic association studies in human alcoholics. Overlapping evidence from animal and human studies will help pinpoint genes and their biological networks that could be most relevant in the treatment of AUD. Mouse drinking models could then be used to test viable targets using approved drugs known to modulate the target network.

Pro-inflammatory neuroimmune signaling pathways have been increasingly implicated in the etiology and progression of AUD and offer unexplored avenues in its treatment. Currently, this is an emphasis area of the Integrative Neuroscience Initiative on Alcoholism (INIA) sponsored by the National Institutes of Health, National Institute on Alcohol Abuse and Alcoholism. Several FDA-approved drugs with anti-inflammatory and immune inhibitory effects regulate alcohol responses in animal models, lending support that some of these could potentially be repurposed to treat AUD. For example, the peroxisome proliferator-activated receptor (PPAR) agonists, fenofibrate and tesaglitizar, reduced drinking in a PPARα-dependent manner (Blednov, Black, Benavidez, Stamatakis, & Harris, 2016), altered expression of immune genes in the liver, and altered neuronal gene expression in mouse brain (Ferguson, Most, Blednov, & Harris, 2014). There is overlapping evidence in mice and humans for specific PPAR genes in alcohol consumption and dependence (Blednov, Benavidez, Black, Ferguson, et al., 2015). In addition, minocycline reduced DID in adult mice, and a bioinformatics pathway analysis revealed an overexpression of neuroimmune-related pathways in these mice (Agrawal et al., 2014). Mouse models will continue to be essential for identifying candidate genes and screening medications to modulate the affected pathways.

Future Directions

While the study of individual genes is informative, combining gene network and systems biology approaches to identify inter-related networks and pathways is critical in the future treatment of AUD. Because complex trait diseases involve coordinated expression changes in multiple gene families, examining gene clusters is an important research direction, as supported by the Integrative Neuroscience Initiative on Alcoholism (INIA) studies showing that co-expression patterns can distinguish gene modules related to alcohol consumption in animal models (Iancu et al., 2013; Nunez et al., 2013; Saba et al., 2011; Saba et al., 2015). Another INIA study reported that RNA-Seq profiling of postmortem human prefrontal cortex also revealed disrupted gene networks in alcohol-dependent individuals compared to matched controls (Farris, Arasappan, Hunicke-Smith, Harris, & Mayfield, 2015). The subnetworks related to lifetime alcohol consumption in humans contained known alcohol targets, from the preclinical literature, and were overrepresented for genes involved in synaptic function. These genes were also enriched in a transcriptome meta-analysis of mouse drinking behavior (Mulligan et al., 2006), suggesting an overlapping set of alcohol-related genes across species. Comparison of human and animal model networks, combined with single gene-based approaches, will be important in understanding the neurobiology of AUD (Farris, Pietrzykowski, et al., 2015).

Genomic and proteomic techniques are capable of analyzing complex data sets into related biological processes (Gorini, Bell, & Mayfield, 2011; Gorini, Harris, & Mayfield, 2014; Gorini, Roberts, & Mayfield, 2013; Sikela et al., 2006). Ideally, integration of genetics and transcriptomics with convergent biological processes and phenotypic behaviors will reveal prospective therapeutic targets for the disease (Gorini et al., 2011; Gorini et al., 2014). Drug addiction is mediated by both genetic and environmental determinants and neuroadaptations in reward circuitry are related to disease progression. Therefore, in addition to individual candidate genes, future research must consider the gene and protein networks and associated biological system changes observed at different stages of the disease. These network-centric approaches can summarize complex gene lists into interrelated components and provide a systems-level picture of the disease. This biological systems framework will be key in determining causal factors of AUD and providing a more integrated, functional approach to its treatment.