Effects of naltrexone on alcohol, sucrose, and saccharin binge-like drinking in C57BL/6J mice: a study with a multiple bottle choice procedure

Chronic alcohol (ethyl alcohol, EtOH) binging has been associated with long-term neural adaptations that lead to the development of addiction. Many of the neurobiological features of EtOH abuse are shared with other forms of binging, like pathological feeding. The drinking-in-the-dark (DID) paradigm has been used extensively to study the neurobiology of EtOH binge-like drinking due to its ability to promote high intakes relevant to human behavior. DID can also generate high consumption of other tastants, but this procedure has not been fully adapted to study forms of binging behavior that are not alcohol-driven. In the present study, we used a modified version of DID that uses multiple bottle availability to promote even higher levels of EtOH drinking in male C57BL/6J mice and allows a thorough investigation of tastant preferences. We assessed whether administration of systemic naltrexone could reduce binging on EtOH, sucrose, and saccharin separately as well as in combination. Our multiple bottle DID procedure resulted in heightened levels of consumption compared with previously reported data using this task. We found that administration of the opioid receptor antagonist naltrexone reduced intakes of preferred, highly concentrated EtOH, sucrose, and saccharin. We also report that naltrexone was able to reduce overall intakes when animals were allowed to self-administer EtOH, sucrose, or saccharin in combination. Our modified DID procedure provides a novel approach to study binging behavior that extends beyond EtOH to other tastants (i.e. sucrose and artificial sweeteners), and has implications for the study of the neuropharmacology of binge drinking.


Introduction
Obsessive cravings and compulsive intake, often in the face of severe personal and medical consequences, are characteristics of drug abuse and binge eating disorders (Rehm et al., 2009;Sacks et al., 2015). In the case of alcohol (ethyl alcohol, EtOH), excessive drinking typically occurs in a binge-like manner (Esser et al., 2014;Kanny et al., 2018), where enough EtOH is consumed to reach high blood EtOH concentrations (BECs) in relatively short periods of time (National Institute on Alcohol Abuse and Alcoholism, 2004). Rates of binge-drinking are high, with 50-90% of all EtOH consumed taking the form of binge episodes in adults and underage adolescents, respectively (Patrick and Schulenberg, 2013;Goings et al., 2019). Binge drinking is associated with accidental injury (Gonzales et al., 2014;Stahre et al., 2014), high blood pressure (Hayibor et al., 2019), increased risk for stroke (Sundell et al., 2008), type-2 diabetes (Pietraszek et al., 2010), and liver dysfunction (Rosoff et al., 2019). Recurrently engaging in binge drinking may also induce long-term neuroadaptations that further promote binging and can lead to the development of addiction (Melendez, 2011;Sprow and Thiele, 2012;Carnicella et al., 2014;Tavolacci et al., 2019). Better understanding of the neurobiological consequences that arise from repeated intermittent exposure to EtOH via binging can lead to improved interventions and outcomes for addiction.
Animal models of EtOH drinking have been developed in an effort to mimic key features of human consumption. Some models have faced criticism due to lack of voluntary drinking and low levels of consumption that do not reach pharmacologically significant BECs (Rhodes et al., 2005;Thiele and Navarro, 2014). One model known as drinking-in-the-dark (DID) takes advantage of rodents' natural circadian rhythms in order to promote high levels of voluntary EtOH drinking (Rhodes et al., 2005). Using this methodology, EtOH-preferring mice consume enough EtOH to show behavioral and pharmacological signs of intoxication that are relevant to human behavior, illustrating the face validity of this procedure (Rhodes et al., 2005;Thiele and Navarro, 2014;Jeanblanc et al., 2019). Although DID has primarily been used to study the neurobiology of binge-like EtOH drinking, it can also produce elevated consumption of other drugs of abuse, including methamphetamine (Fultz et al., 2017), opioids (Szumlinski et al., 2019), and other tastants like sugar (Kamdar et al., 2007;Cozzoli et al., 2012;Giardino and Ryabinin, 2013;Holgate et al., 2017). Although a more thorough investigation of the conditions that elicit binging on sweet tastants has not been conducted, the ability of DID to promote elevated intake points at this paradigm's potential for understanding common mechanisms underlying binging on EtOH and other palatable rewards.
Alcohol abuse has been proposed to share many neurobiological mechanisms with other pathological forms of consumption, including high-calorie sweet foods (Avena, et al., 2012a,b;Schulte et al., 2015Schulte et al., , 2016. A strong correlation between high levels of EtOH intake and sweet food consumption has been shown in humans (Kampov-Polevoy et al., 1999;Leggio et al., 2011). Preference for strong sweet solutions has also been shown to be associated with a paternal history of alcohol dependence (Kampov-Polevoy et al., 2001. This history of dependence has additionally been identified as a significant predictor of 'sugar-addiction' in children, suggesting a close link between alcoholism and high sweet preference with some heritable aspects (Fortuna, 2010). Similar findings have been observed in rodents and other preclinical models. For example, strains of mice that voluntarily consume high levels of EtOH such as C57BL/6J (B6) show higher preference indices for sucrose in comparison to other strains (Bachmanov et al., 1996(Bachmanov et al., , 2001. Rats characterized as 'sugar-dependent' show enhanced intake of unsweetened EtOH solutions (Avena et al., 2004). Mice sensitized to the psychomotor effects of EtOH display altered patterns of sucrose consumption, showing a more rapid initial approach and consumption of sucrose in EtOH-sensitized groups, suggesting that EtOH-induced neuroplasticity can affect consummatory behaviors for sweet rewards (Pastor et al., 2010).
EtOH and sweet tastants like sugar share chemosensory mechanisms of action (Di Lorenzo et al., 1986;Scinska et al., 2000) and significantly overlap in their actions on brain reward systems (Bodnar, 2019;Olszewski et al., 2019), including the endogenous opioid system. Rewards like EtOH and sugar may exert powerful effects on pleasure 'liking' via coordinated actions on brain opioid receptors within a network of hedonic 'hotspots' (Berridge and Kringelbach, 2015;Castro and Berridge, 2017). Excessive binging on sugar or EtOH, to the point of generating signs of dependence, can produce negative affective states of withdrawal that are opioid-dependent (Colantuoni et al., 2002;Avena et al., 2009;Berger et al., 2013). Opioid receptors, and in particular the mu-receptor can also influence the motivational and psychomotor stimulant effects of EtOH and other reinforcers (Gianoulakis, 2001;Pastor et al., 2005Pastor et al., , 2011Pastor and Aragon, 2006;Kamdar et al., 2007). Pharmacological activation of mu-opioid receptors can enhance saccharin or EtOH intakes , while inactivation can reduce consumption of sucrose, saccharin, or EtOH (June et al., 2004;Kamdar et al., 2007;Tarragón et al., 2012;Avena et al., 2014;Morales et al., 2017). The opioid receptor antagonist naltrexone, a compound with high affinity for mu-opioid receptors, is used to treat alcoholism (Volpicelli et al., 1992;Kiefer et al., 2003;Jonas et al., 2014), and in combination with bupropion (marketed as Contrave, Therapeutics, Inc.; La Jolla, California, USA) has recently been approved for the treatment of binge-eating disorder and weight-loss management. With B6 mice and a dose range of 2-8 mg/ kg, previous data have shown that naltrexone can reduce EtOH binge-like intake using the DID model, although some sex differences have been reported (Kamdar et al., 2007;Tarragón et al., 2012;Navarro et al., 2019;Zhou et al., 2019 ). It is important to mention, however, that doses of naltrexone as high as 10 mg/kg have failed to reduce EtOH intake in mice selected for high blood EtOH levels [high drinking-in-the-dark (HDID lines)] obtained using the DID model (Crabbe et al., 2017). HDID lines show remarkably high EtOH intakes (Crabbe et al., 2014). Higher doses of naltrexone might be therefore needed to reduce high binge-like EtOH intakes. Overall, although more research is needed, cumulative evidence on the effects of naltrexone on both EtOH and palatable food binging suggests that this compound can be a highly useful tool to explore the predictive validity of a preclinical model.
The present study was designed to investigate whether a modified multiple bottle DID paradigm could be used to study similarities in the neural alterations that result from binging on EtOH, sucrose, and saccharin. This concurrent multiple bottle (three or four bottles) procedure has several advantages over previous one-bottle and two-bottle iterations. Concurrent access to multiple bottles, whether in the same or different concentrations can produce magnified intakes (Serra et al., 2003;Tordoff and Bachmanov, 2003a;Bell et al., 2006;Rodd et al., 2009;Cozzoli et al., 2012;Colombo et al., 2014;Fultz et al., 2017), so this paradigm has the capacity to increase the likelihood of binging behavior. In addition, it also allows for the investigation of the effects of various concentrations on preference and intake. We used the modified DID model to show elevated consumption of EtOH and other powerful natural reinforcers. Given that the effects of naltrexone have been consistently shown to reduce binge-like behavior, we administered systemic naltrexone in order to determine whether it could not only reduce binge-like consumption of EtOH, sucrose, and saccharin but also alter preference for various concentrations.
were individually housed (with no enriched environment) and kept on a standard 12-h light/dark cycle with lights on at 8:00 a.m. Colony rooms were maintained at 21 ± 1°C of temperature and 50 ± 5% humidity levels. Food and water were provided ad libitum throughout the study unless otherwise indicated. Mice were acclimated to housing and colony conditions for 2 weeks before experiment initiation. All experiments were conducted in accordance with the guidelines provided by the European Community Council Directive (2010/63/EU) for the use of laboratory animal subjects and approved by the Animal Care Committee of Universitat Jaume I.

Experimental design
To evaluate binge-like drinking, we used a modified version of the 4-day DID procedure introduced by Rhodes et al. (2005) in which B6 mice drink EtOH voluntarily to the point of behavioral intoxication (Rhodes et al., 2005Kamdar et al., 2007, Lowery et al., 2010Fultz et al., 2017). Three hours after the start of the dark cycle, water tubes were replaced with 10 mL graduated (0.1 mL increments) cylinders with double-ball bearing sipper tubes containing different solutions depending on the experiment. For each experiment (except for experiment 1), four drinking cycles (one per week) of 4 days were used. During the first 3 days of each cycle, drinking tubes were available for 2 h, then replaced with home cage water bottles. On day 4 (test day), tubes were available for 4 h, then replaced with water. Food was not removed when EtOH, sucrose, or saccharin tubes were introduced. Days 1-3 provided habituation to the procedure. On day 4 (Experiments 2-6; intakes recorded at 2 and 4 h), animals received intraperitoneal injections of naltrexone (0, 4, 8, or 16 mg/kg) 30 min before water tubes were replaced with drinking solutions. We used a counterbalanced, within-subject design; all subjects received all doses of naltrexone, one per week (avoiding ascending or descending dose schedules). For each cycle, habituation was conducted Tuesday-Thursday, with test on Friday. Animals were left undisturbed (with regular food/ water availability) for 3 days a week; Saturday-Monday. No significant intake differences between cycles (weeks) were found. Body weights were recorded on test days. Rack-mounted empty cages with solution tubes were used as a control and correction for leakage. With double-ball bearing sipper tubes leaks were rare and minimal (i.e. 0.1 mL).
Experiment 1 evaluated DID intake using different EtOH concentrations and tube availability configurations (n = 12-15 per group). For this experiment, drinking was only assessed for one cycle. Following previous results (Cozzoli et al., 2012;Fultz et al., 2017), single and multiple tube configurations were used with different groups.
EtOH 20 and 40% were tested using one single tube. Using three simultaneously available tubes, animals were offered EtOH 5, 10, and 20%, and also EtOH 20, 30, and 40%. Using four concurrently available tubes, animals had access to EtOH 5, 10, 20, and 40%. This experiment allowed us to select a multiple-tube choice version of the DID model that produced high EtOH intake.
Given that we wanted to evaluate the effects of naltrexone on intake and concentration preference of EtOH, sucrose, or saccharin, a multiple tube choice procedure was used for all substances. The effects of naltrexone on EtOH, sucrose, or saccharin intake were determined using four simultaneously available tubes. EtOH 5, 10, 20, and 40% (experiment 2, n = 15), sucrose 5, 10, 20, and 40% (experiment 3, n = 15), or saccharin 0.13, 0.26, 0.53, or 1.06% tubes (experiment 4, n = 16) were placed using a counterbalanced order of concentrations that was changed every day to avoid prediction of tube placement. Complete ascending or descending arrangements of concentration tubes were also avoided. A separate experiment (experiment 5; n = 9) to test the effect of naltrexone on water intake was conducted. This experiment followed the same design (four cycles of a 4-day DID procedure) and doses described for Experiments 1-3, except that four drinking tubes containing water were used. In experiment 6 (n = 15), we were interested in evaluating the effects of naltrexone on intake using a procedure involving simultaneously available sucrose, saccharin and EtOH (three tubes). Animals could choose to drink 5% sucrose, 0.13% saccharin, or 20% EtOH. Given the strong preference for sweet solutions, the two lowest sucrose and saccharin solutions were used against the preferred EtOH concentration (20% EtOH; based on experiment 2 results; milliliters).

Data analysis
Data were analyzed using GraphPad Prism 8 software (GraphPad Software Inc., San Diego, California, USA). Time period and naltrexone dose were the independent variables. The dependent variables were intakes of EtOH, sucrose, or saccharin (milliliters and grams per kilogram) in Experiments 1-4, intake of water (milliliters) in experiment 5, and intake of sucrose, saccharin plus EtOH (milliliters and milliliters per gram) in experiment 6. In experiment 5 all tubes contained water; water intake was therefore only analyzed as total milliliters drunk. Due to the substantial differences in grams of EtOH, sucrose, or saccharin contained per milliliter of these substances, intakes adjusted per body weight in experiment 6 were expressed as milliliters per gram, instead of grams per kilogram. As described before (Morales et al., 2017), given that we used a within-subjects pharmacological design, the effects of naltrexone were analyzed using repeated measures analysis of variance (ANOVA). Apart from time (first 2 h vs. second 2 h), different simultaneously available concentrations or solutions were also considered repeated factors. Experiment 1 was analyzed using a one-way ANOVA. Post-hoc comparisons were conducted using using Tukey's Honest Significant Difference tests. The level for significance (α) for all statistical tests was set at 0.05.

Results
Intake as a function of ethyl alcohol concentration and tube availability: comparison to single-tube drinking-in-the-dark Figure 1 shows EtOH intakes (grams per kilogram; 4 h) of separate groups of animals that were offered EtOH under five different tube/concentration configurations. A one-way ANOVA (F 4,61 = 18.03, P < 0.01) revealed a significant effect of tube condition. From the two single-tube groups (A and B), those that had access to 40% EtOH showed increased intake EtOH (P < 0.05, compared with 20% EtOH). The two groups with the highest concentrations available (D and E), with three (20, 30, and 40%) and four (5, 10, 20, and 40%) tubes, respectively, also showed the highest intakes (P < 0.01, compared with 20% EtOH). Group E was also different from group B (P < 0.05) and showed the highest intake during the first 2 h (data not shown) of the 4-h period. Respect to intakes per concentration in multiple tube groups, with concurrently available 5, 10, and 20% EtOH tubes (group C), most of their intake came from the 20% tube (3.63 g/kg; 71.6% of total intake). With 20, 30, and 40% EtOH (group D), intake from the three different tubes was similar; 2.58 g/kg from EtOH 20%, 3.03 g/kg from EtOH 30%, and 2.95 g/kg from EtOH 40%, representing 30.14, 35.40, and 34.46% of total intake, respectively. With four tubes (group E), animals obtained most of the grams per kilogram from EtOH 20% (3.3 g/kg; 36.7%) and EtOH 40% (4.77 g/kg; 52.5%). Table 1 shows the effects of naltrexone on volumes of EtOH drinking (milliliters). We first assessed whether naltrexone reduced total EtOH consumption by combining all EtOH concentrations. Over the entire 4-h period (top panel), a one-way ANOVA revealed that naltrexone reduced EtOH drinking (F 3,42 = 10.7, P < 0.01). Both 8 mg/ kg (P < 0.05) and 16 mg/kg (P < 0.01) suppressed EtOH intake relative to saline, although 16 mg/kg also differed from 4 mg/kg (P < 0.01). We also assessed whether naltrexone differentially suppressed total intakes during the first and second 2-h periods. Using a two-way ANOVA (dose × time), we found main effects of time (F 1,14 = 5.57, P < 0.05) and naltrexone dose (F 3,42 = 10.35, P < 0.01) with no interaction, suggesting that EtOH drinking was higher in the first 2-h period and naltrexone suppressed the volume of EtOH drank. We further assessed whether naltrexone altered drinking as a function of concentration. For the first 2 h of intake, a two-way ANOVA (concentration × dose) revealed significant effects of concentration (F 3,42 = 14.76, P < 0.01), naltrexone dose (F 3,42 = 7.85, P < 0.01), and an interaction between factors (F 9,126 = 2.4, P < 0.05). Under baseline conditions, mice equally sampled between 40 and 20% EtOH, which were both preferred to the 5% option (P < 0.01). However, 8 mg/ kg (P < 0.01, compared with naltrexone 0) and 16 mg/kg (P < 0.01, compared with 0 and 4 mg/kg) naltrexone treatments selectively reduced this preference by suppressing intake of the 20% solution. We found similar results over the second 2-h period, with significant effects of EtOH intake (grams per kilogram) as a function of EtOH concentration availability. Intakes obtained with separate groups (n = 12-15) with access (4 h) to one, three or four EtOH tubes. From left to right, bars (mean ± SEM) represent access to: (a) one tube with 20% EtOH; (b) one tube with 40% EtOH; (c) three simultaneously available tubes with 5, 10, and 20% EtOH; (d) three simultaneously available tubes with 20, 30, and 40 EtOH; (e) four simultaneously available tubes with 5, 10, 20, and 40% EtOH. *P < 0.05, **P < 0.01 indicates different from (a). Group (e) was also found to be different from (b) (P < 0.05). EtOH, ethyl alcohol. concentration (F 3,42 = 8.97, P < 0.01), dose (F 3,42 = 5.66, P < 0.01), and interaction between factors (F 9,126 = 3.49, P < 0.05). However, unlike the first 2-h period, during which both 40 and 20% were sampled, 20% was the most preferred concentration of EtOH (P < 0.01 relative to 5 and 10%). The highest naltrexone dose abolished this Top panel shows the effects of naltrexone (0, 4, 8, or 16 mg/kg) on total EtOH intake (milliliters) collapsed across different concentrations for the 0-2, 2-4, and 0-4 h time periods; *P < 0.05 different from 0 mg/kg, **P < 0.01 different from 0-4 mg/kg. Bottom panels show the effects of naltrexone on EtOH drunk from each individual tube (concentration); *P < 0.05, **P < 0.01 relative to 0 mg/kg for each respective time period and concentration. For the 20% EtOH, 16 mg/kg naltrexone groups **P < 0.01 also indicates different from 4 mg/kg (2-4 and 0-4 h time periods). EtOH, ethyl alcohol.

Fig. 2
Effects of naltrexone on EtOH intake (grams per kilogram). Panel A shows the effect of naltrexone (0, 4, 8, or 16 mg/kg) on total EtOH intake (collapsed on concentration) during 0-2, 2-4, and 0-4 h time periods. Panels B-D show naltrexone effects on EtOH intake as a function of concentration for the two 2-h periods and total 0-4 h period, respectively. *P < 0.05, **P < 0.01 indicates different from naltrexone 0 mg/kg at the same EtOH concentration. Bars represent mean ± SEM; n = 15 per group. EtOH, ethyl alcohol.

Fig. 3
Effects of naltrexone on sucrose intake (grams per kilogram). Effects of naltrexone for total intakes during 0-2, 2-4, and 0-4 h time periods, combining all sucrose concentrations, are shown on panel a; *P < 0.05, different from naltrexone 0 mg/kg for each respective time period. Panels b-d show naltrexone effects on sucrose intake as a function of concentration during the two 2-h periods and total 0-4 h period, respectively. **P < 0.01 indicates different from naltrexone 0 mg/kg at the same EtOH concentration. Bars represent mean ± SEM; n = 15 per group. EtOH, ethyl alcohol. 16 mg/kg were different from saline (P < 0.01), and naltrexone 8 mg/kg was also different from naltrexone 4 mg/ kg (P < 0.05).

Effects of naltrexone on water intake
Naltrexone administration did not affect water consumption (milliliters). With data including the two separate 2-h periods, a two-way ANOVA (time period × naltrexone dose) revealed an effect of time (F 1,8 = 8.76, P < 0.05) but no effects of naltrexone or interaction between factors. Naltrexone was also found to produce no effects on water intake when total 4-h data were analyzed. Group means (mean milliliters ± SEM) during the first 2 h were 0.67 ± 0.08, 0.74 ± 0.07, 0.67 ± 0.09, and 0.61 ± 0.06 (naltrexone 0, 4,

Effects of naltrexone on simultaneously available sucrose, saccharin, and EtOH consumption
The top panel (Table 4) presents combined intakes from the three tubes (milliliters) during the two separate 2-h periods as well as total 4-h period. We found a main effect Effects of naltrexone on saccharin intake (grams per kilogram). Panel a shows the effect of naltrexone on total intake (combining all saccharin concentrations) during the 0-2, 2-4, and 0-4 h time periods; *P < 0.05, different from naltrexone 0 mg/kg for each respective time period. Panels b-d show naltrexone effects on saccharin intake at different concentrations and the effects of naltrexone during the two 2-h periods and total 0-4 h period, respectively. *P < 0.05, **P < 0.01 indicates different from naltrexone 0 mg/kg at the same EtOH concentration. Bars represent mean ± SEM; n = 16 per group. EtOH, ethyl alcohol. of naltrexone (F 3,42 = 7.03, P < 0.01) on combined drinking but there was no time × naltrexone interaction. Total volume consumption over 4-h testing was reduced following administration of all naltrexone doses (P < 0.01 for naltrexone 4 and 8, and P < 0.05 for naltrexone 16 mg/ kg). We then analyzed whether naltrexone treatment selectively affected drinking of the various tastants using a two-way ANOVA (dose × tastant). An effect of tastant was found for the first (F 2,28 = 137.10, P < 0.01) and second (F 2,28 = 146.40, P < 0.01) time periods. For the second 2-h period, an effect of naltrexone was also found (F 3,42 = 10.26, P < 0.01). No interactions between factors were found for both periods. Analysis of total 4-h intakes with the three separate tastants (dose × tastant) showed effects of tastant (F 2,28 = 182.70, P < 0.01) and naltrexone (F 3,42 = 7.03, P < 0.01), but no interaction. With data adjusted to body weight ( Fig. 5; milliliters per gram), analysis of intakes combining all solutions ( Fig. 5a; dose × time) showed an effect of naltrexone (F 3,42 = 6.35, P < 0.01), but no effects of time or dose × time interaction. Pairwise tests (0-4 h) showed that all naltrexone doses reduced total intake; P < 0.05 for naltrexone 4 and 16, and P < 0.01 for naltrexone 8 mg/kg. When the three different solutions were analyzed separately we found an effect of tastant for both, first (F 2,28 = 109.40, P < 0.01) and second (F 2,28 = 140.10, P < 0.01) 2-h periods ( Fig. 5b and c, respectively) but no significant interactions between naltrexone and tastant. An effect of naltrexone was also found on the second period (F 3,42 = 11.74, P < 0.01). Similarly, main effects of drinking solution (F 2,28 = 166.11, P < 0.01) and naltrexone (F 3,42 = 6.35, P < 0.01) were found for the total 4-h period (Fig. 5d), without a significant interaction.

Discussion
Understanding the brain mechanisms by which vulnerable individuals develop unmanageable patterns of consumption of food or alcohol is key in order to improve treatments and prevent addiction. Over the last decade, an increasing number of studies have used animal models that attempt to mimic binge intake (Thiele and Navarro, 2014;Jeanblanc et al., 2019;Treasure and Eid, 2019). Originally designed to study high voluntary EtOH consumption achieving behaviorally-intoxicating BECs, the DID model has been extensively used to investigate the neurobiology of binge-like EtOH intake (Rhodes et al., 2005;Thiele and Navarro, 2014). In the present study, we introduced a modified version of the DID procedure that further magnifies EtOH intake by the use of multiple bottles. This procedure was also shown to induce high sucrose and saccharin consumption. Our Effects of naltrexone on simultaneously available saccharin (S), sucrose (SU), and EtOH (E) intake (milliliters per gram). Effects of naltrexone on total intakes during 0-2, 2-4 , and 0-4 h time periods, combining milliliters drank from saccharin, sucrose, and EtOH tubes are shown on panel A; *P < 0.05, **P < 0.01 different from naltrexone 0 mg/kg. Panels B-D show naltrexone effects on the three different drinking solutions during the two 2-h periods and total 0-4 h period, respectively. Bars represent mean ± SEM; n = 15 per group. EtOH, ethyl alcohol.
pharmacological data provide new evidence supporting a role of opioid receptors in heightened EtOH and sweet tastant intake, suggesting common neurobiological pathways that may be affected as a function of enhanced consumption.
Although there are several variations of the DID model, one of the most frequently used protocols in EtOH research involves replacing the water bottle with a tube containing 20% EtOH for 2-4 h, beginning 3 h into the dark cycle (Rhodes et al., 2005;Thiele and Navarro, 2014). With the commonly used 4-day version of this protocol, EtOH access is given over 2 h on days 1-3 but extended to a 4-h period on day 4 (test day). Using this DID protocol, different laboratories have demonstrated that genetically predisposed rodent strains such as B6 mice will drink up to ~5-7 g/kg of EtOH in 4 h (Rhodes et al., 2005;Tarragón et al., 2012;Thiele and Navarro, 2014). This level of intake produces BECs higher than 100 mg/ dL and induces clear signs of behavioral intoxication evidenced by motor impairment on the rotarod and balance beam tasks . It is important to mention that the National Institute on Alcohol Abuse and Alcoholism (2004) defines binge drinking in the context of blood alcohol concentrations (≥80 mg/dL). Given the higher rate of metabolism, minimum BECs of 100 mg/ dL BECs have been proposed for mice. Interestingly, Rhodes et al. (2007) demonstrated that while water availability (two-bottle choice test; water vs. 20% EtOH) during a 4-h test did not reduce EtOH intake, it did affect BECs (averaging less than 80 mg/dL, as opposed to BECs greater than 120 mg/dL with access to EtOH alone). A 4-day DID protocol with a 4-h test on day 4 using a single 20% EtOH tube has therefore been extensively used to investigate the biological determinants of binge-like EtOH drinking (Thiele and Navarro, 2014;Jeanblanc et al., 2019 for reviews). In our laboratory, we have previously used a 4-h DID procedure with a single 20% EtOH tube (Tarragón et al., 2012) achieving 5.98 ± 0.29 g/kg of EtOH (male B6 mice). This intake resulted in a range in BECs of 120-140 mg/dL, with a strong positive correlation between grams per kilogram and BECs. Similar to what was described by Rhodes et al. (2005), we found comparable levels of EtOH intake using single 10, 20, or 30% EtOH tubes; milliliters consumed across concentrations changed to achieve similar levels of grams per kilogram (Tarragón et al., 2012).
Previous research has demonstrated that the number of bottles available during a drinking test influences EtOH intake; availability of two or more EtOH bottles of the same or different concentration increases EtOH intake in mice and rats (Spanagel et al., 1996;Serra et al., 2003;Tordoff and Bachmanov, 2003a). This phenomenon has also been described for other taste solutions such as saccharin, citric acid, quinine, and sodium chloride (Tordoff and Bachmanov, 2003b). In the context of binge-like drinking, the use of protocols involving multiple concurrently available EtOH concentrations (2-4 bottles), has also been shown to produce notable increases in EtOH intake in alcohol-preferring (P) and Sardinian alcohol-preferring (sP) rats (Bell et al., 2006;Colombo et al., 2014), as well as in B6 and mixed B6 × 129X1/SvJ (B6×129) mice (Cozzoli et al., 2012;Fultz et al., 2017). Similarly, Shabani et al. (2016) used selectively bred methamphetamine high drinking mice to demonstrate that the availability of three (vs. 2 or 1) bottles of methamphetamine also increased drug intake. These studies suggest that for multiple rewards, greater availability leads to heightened consumption. Considering this evidence, we wanted to investigate whether we could boost binge-like EtOH intake with the concurrent availability of multiple EtOH bottles. Additionally, by presenting different EtOH concentrations simultaneously, we wanted to evaluate whether naltrexone would reduce overall EtOH consumption (Kamdar et al., 2007;Tarragón et al., 2012) or affect the pattern of EtOH intake preference across concentrations. In preparation for this study, our data with B6 mice (experiment 1) showed that a 4-day DID procedure (4-h test on day 4) with concurrently available 5, 10, and 20% EtOH tubes produced an intake of 5.08 ± 0.35 g/kg of EtOH. We also tested concurrent 20, 30, and 40% EtOH, which produced a particularly high level of EtOH intake; 8.39 ± 0.59 g/ kg. In this case, intakes obtained from each tube were comparable, evidencing B6's remarkable preference for highly concentrated EtOH solutions. In fact, with a single 40% EtOH tube animals drank up to 6.87 ± 0.48 g/kg. Favoring larger differences across concentrations (Fultz et al., 2017), for the present study we chose a concurrent 4-bottle DID procedure with 5%, 10, 20, and 40% EtOH. With this protocol B6 mice drank up to 9.05 ± 0.61 g/kg EtOH in 4 h (mostly from the 20 and 40% tubes); drinking approximately 60% of total intake (5.25 ± 0.42 g/kg) during the first 2 h. Compared to our own data using 4-h DID tests (current study and Tarragón et al., 2012), our present intake was >2.5 g/kg higher than what we found with a single 20% EtOH bottle. Earlier studies using DID procedures with multiple bottles have also shown increased intake when compared with single-tube 20% EtOH studies. Cozzoli et al. (2012) demonstrated that a two-bottle DID test (concurrent 5 and 20% EtOH) produced intakes of ~5 g/kg (2 h) in B6×129 mice. Particularly, relevant for the present study are the data presented by Fultz et al. (2017) with the same four concentrations used here; they showed impressive 2-h DID intakes of ~6 g/kg in B6 and ~7.5 g/kg in B6×129 male mice in 2 h. Altogether, these data indicate that multiple bottle choice DID procedures can produce particularly high EtOH intakes. Whether this high drinking behavior is associated with availability of different concentrations, the number of tubes that are presented, or both, will need to be elucidated in future experiments.
It is important to mention that in the present study we did not measure BECs. However, given our previous data (Tarragón et al., 2012) and those published by others (Crabbe et al., 2011), we can definitely suggest that the intakes of EtOH seen here would be expected to exceed 100 mg/dL. It is also relevant to point out that our data reflect intakes averaged over a period of 4 weeks (four, 4-day cycles; 4 days of EtOH availability followed by 3 days with no EtOH exposure). It could be argued that the intakes reported here were influenced by escalated drinking over repeated cycles of drinking and withdrawal. Although this effect has been previously reported in B6 (Fultz et al., 2017), and a trend was seen in our data, we did not find significant differences in overall EtOH intake across weeks 1-4 (8.34 ± 1.04 and 9.74 ± 1.32 g/kg for weeks 1 and 4, respectively; t-test, P > 0.05). It might be possible that longer periods of time are required to find clear time-dependent differences. Wilcox et al. (2014) found changes in EtOH consumption when DID was extended over a period of 6 weeks; intakes of weeks 4-6 were higher than those recorded during weeks 1-3. Additionally, this study revealed an interesting pattern of EtOH intake (described as 'front-loading' behavior) characterized by an increased rate of consumption during the first 15 min of the drinking session. The amount of EtOH drunk during the first part (first 30 min) of the session was also found to double from week 1 to week 6, indicating that this phenomenon developed over time (Wilcox et al., 2014). The use of repeated cycles of DID including multiple EtOH bottles might be an interesting strategy to boost binge-level drinking in future studies.
The predictive validity of DID as an animal model has been tested using drugs approved to treat alcoholism, including the opioid receptor antagonist naltrexone. Naltrexone is known for its efficacy in reducing high alcohol drinking in relapsing patients (Volpicelli et al., 1992;Kiefer et al., 2003;Jonas et al., 2014). Naltrexone, alone or in combination with other drugs, has been shown to reduce EtOH intake using DID methodology (Kamdar et al., 2007;Tarragón et al., 2012;Ripley et al., 2015;Zhou et al., 2019;Navarro et al., 2019). These results add to an extensive list of studies indicating that naltrexone reduces EtOH consumption using a variety of methodologies and animal models (Davidson and Amit, 1997;Sharpe and Samson 2001;Fachin-Scheit et al., 2006), including genetically predisposed EtOHpreferring strains of rodents such as B6 mice (Lê et al., 1993;Phillips et al., 1997;Middaugh et al., 2003;Dhaher et al., 2012). The current study presents new evidence showing that high binge-like EtOH drinking can be reduced by naltrexone. In particular, a reduction in intake was seen at the doses of naltrexone 8 and 16 mg/ kg, but not 4 mg/kg. We have previously shown reductions in EtOH drinking with naltrexone 4 mg/kg using DID tests with a single 20% EtOH tube (Tarragón et al., 2012), an effect also found by Kamdar et al. (2007) with naltrexone 2 mg/kg. The particularly high intake of EtOH found with our procedure may be associated with the fact that higher doses of naltrexone were needed to reduce EtOH consumption. Consistent with this, Crabbe et al. (2017) showed that naltrexone 10 mg/kg failed to reduce EtOH drinking in mice selectively bred for their high BECs using a DID procedure (HDID mice), which also showed high EtOH intake. Additionally, we observed a stronger, dose-dependent naltrexone effect during the first 2 h; naltrexone reduced consumption of those concentrations that capitalized intake (EtOH 20 and 40%). However, during the second 2 h, an increased preference for the highest EtOH concentration diminished the overall effect of naltrexone. As a result, for the total 4-h period, a reduced 20%, but not 40% EtOH intake was found with naltrexone 16 mg/kg. By increasing intake from the most pharmacologically efficient EtOH tube (40%), B6 mice might have tried to oppose the reward-reducing effects of a high naltrexone dose. Overall, the present DID data support extensive evidence proposing a role of the endogenous opioid system in the neurobiology that mediates EtOH drinking (Gianoulakis, 2001;Font et al., 2013). Future studies will need to elucidate the involvement of mu-, delta-, and kappa-opioid receptors using EtOH DID protocols. Mice lacking the proopiomelanocortin-derived peptide beta-endorphin showed reduced DID using 7.5, 15, or 30% EtOH solutions, suggesting a pivotal role of muand delta-opioid receptors (Zhou et al., 2017a). The endogenous opioid system has also been linked to sugar and other sweet rewards Olszewski and Levine, 2007;Berridge and Kringelbach, 2015). Intake of saccharin solutions increases by infusions of a mu-opioid agonist into the ventral striatum , and pharmacological antagonism of opioid receptors can reduce consumption of sucrose and saccharin (Biggs and Myers, 1998;June et al., 2004;Morales et al., 2017). Also, excessive sugar intake can produce withdrawal-like symptoms associated with alterations in opioid receptor signaling (Colantuoni et al., 2002;Avena et al., 2009). In our study, B6 mice showed remarkably high intakes of sucrose, drinking up to 32.75 ± 1.40 g/ kg in 4 h (0-2 h; 19.5 ± 1.39 g/kg, and 2-4 h; 13.25 ± 1.10 g/ kg). Our previous data using a 2-h 10% sucrose DID test showed that B6 mice consumed ~6.8-8.2 g/kg (Tarragón et al., 2012). In our current study, intake (both in milliliters and grams per kilogram) was obtained almost exclusively from the 40% tube. Our pilot studies (data not shown) revealed that, regardless of the combination of concentrations, mice always obtained most of their intake from the highest sucrose concentration available. Due to extreme solution viscosity, we did not use concentrations higher than 40%. With our procedure, we found that the two highest doses of naltrexone (8 and 16 mg/kg) reduced sucrose drinking, an effect mostly seen during the first 2 h of the 4-h test. Studies using less concentrated solutions (i.e. 10%) and lower doses of naltrexone have reported no effects of this antagonist on DID sucrose consumption (Kamdar et al., 2007;Tarragón et al., 2012). Future studies will need to clarify whether, as the current data suggest, the effect of naltrexone on sucrose DID requires higher levels of binge-like intake and/or higher doses of naltrexone.
High saccharin intake was also achieved with our multiple-bottle DID protocol, reaching 0.72 ± 0.05 g/kg in 4 h (0-2 h; 0.42 ± 0.04 g/kg, and 2-4 h; 0.30 ± 0.03 g/kg). This intake was significantly higher than what has been previously reported with other DID experiments that used lower concentrations. Using 0.1% saccharin, Zhou et al. (2019) found intakes around 0.1 g/kg in 4 h (male B6), while Kamens et al. (2018) showed drinking up to ~0.02 g/kg in 2 h (adolescent male and female B6; 0.033% saccharin). Our animals preferred the concentration of 1.06%. However, this preference for the highest concentration was not as skewed as that observed with sucrose. Approximately 50 and 20% of total intake (milliliters) was obtained from the 1.06 and 0.53% tubes, respectively. To the best of our knowledge, the current study presents the first data showing that naltrexone reduces saccharin intake using a DID procedure. A combination of bupropion 10 mg/kg + naltrexone 1 mg/kg, capable of reducing 4% sucrose intake, did not affect 0.1% saccharin DID consumption (Zhou et al., 2019). However, previous research using non-DID procedures has found that opioid receptor antagonism can reduce saccharin intake (Lynch and Libby, 1983;Chow et al., 1997;Biggs and Myers, 1998). In our case, high saccharin DID was particularly resistant to the effects of naltrexone; we had to administer 16 mg/kg of naltrexone to see a reduction in overall intake. Similar to what we found with sucrose, this effect was mostly observed during the first 2 h. Concentration-dependent analyses showed that naltrexone 8 mg/kg also reduced intake of 1.06% during the first 2 h. Interestingly, during the 2-4 h period, naltrexone 16 mg/kg increased intake of saccharin 0.53% (in milliliters) and did not reduce 1.06% drinking. This effect mirrors our findings in the EtOH study and might reflect compensatory mechanisms aimed to outweigh the effects of naltrexone.
One advantage of multiple-bottle drinking procedures is that they can also be used to assess intake preferences across different solutions. In the present study, we investigated consumption of low concentrations of sucrose (5%) and saccharin (0.13%) concurrently available with 20% EtOH. Mice preferred sweet solutions over EtOH, overall consuming most from the saccharine option, followed by sucrose. In comparison to our single tastant studies, total intakes were significantly higher for this experiment, with values exceeding 4.5 mL in 4 h. Perhaps the fact that sucrose and saccharin were offered at lower concentrations allowed them to drink more volume. This elevated drinking prompted us to examine water intake during the remaining 20 h (nonexperimental phase), finding that mice consumed ~90% of their daily fluid intake during the 4-h DID test (data not shown). It is also interesting to point out that, with this combination of solutions, total intakes during the first and second 2-h periods were comparable. This differs from our single solution experiments where animals tended to 'front-load' intakes in the first 2-h period. In addition, the effect of naltrexone on intake with three different solutions was also different. We found that all naltrexone doses, including 4 mg/ kg, reduced intake (overall 4-h drinking), and this effect was particularly significant during the second 2 h of the test. Sucrose and saccharin intakes were not differentially affected by naltrexone. This is important to highlight because this experiment was originally designed to evaluate whether naltrexone would produce differential effects on intake when animals can choose between caloric (EtOH, sucrose) and noncaloric (saccharin) solutions. Our data suggest that both, taste and postingestive aspects of sucrose and saccharin intake can be regulated by opioid receptors. It has been argued that homeostatic factors can affect DID (Tabarin et al. 2007;Thiele and Navarro, 2014). However, manipulations to the homeostatic state of animals, either by food deprivation or pharmacological manipulations of molecules that influence appetite have all failed to alter the high levels of EtOH intake produced by the DID procedure (Lyons et al., 2008). This suggests that such levels of consumption are likely to be driven by the reinforcing pharmacological effects of EtOH. The ability of naltrexone to reduce intake is therefore likely due to its ability to interfere with such reinforcing properties of EtOH. Future experiments with food-deprived and sated animals might help clarify the role of the opioid system in modulating sucrose and saccharin binge-like drinking using DID methodology.

Conclusions and future directions
Drug abuse and binge eating disorders share imbalances in brain systems that regulate motivation, reward saliency, decision-making, and self-control (Volkow et al., 2017;Wiss et al., 2018). Together with mesolimbic dopamine, the endogenous opioid system has been shown to play an important role in regulating food and drug reward (Volkow et al., 2013(Volkow et al., , 2017Berridge and Kringelbach, 2015). Clinical evidence indicates that endogenous opioids promote intake of sweet and fat tastants, and a polymorphism of the mu opioid receptor gene has been associated with binge eating disorders (Davis et al., 2009;Olszewski et al., 2011;Volkow et al., 2013Volkow et al., , 2017. In the current study we introduced a modified version of an animal model of binge-like intake (DID) that produces notably high consumption of EtOH, sucrose and saccharin, and present evidence of an involvement of opioid receptors in such intakes. This high binge-like intake-promoting procedure expands our methodological options to investigate overconsumption of palatable foods and addictive substances. In future studies, it could be used to investigate the role of sex, energy homeostasis, history of drug experience as well as to test new pharmacological tools in the context of binge intake research.