Home Cage Voluntary Alcohol Consumption Increases Binge Drinking without Affecting Abstinence-Related Depressive-Like Behaviors or Operant Responding in Crossed High Alcohol-Preferring (cHAPs)

Abstract

Chronic alcohol consumption can lead to tolerance and escalation of drinking in humans and animals, but mechanisms underlying these changes are not fully characterized. Preclinical models can delineate which mechanisms are involved. The chronic intermittent ethanol exposure (CIE) procedure uses forced exposure to vaporized alcohol that elicits withdrawal and increased responding for alcohol in operant tasks in C57BL/6J inbred mice (Lopez et al, 2008; Kliethermes et al., 2006). Chronic two-bottle choice (2BC) drinking in the same strain elicits abstinent-related depression-like behavior (Stevenson et al. 2009), suggestive of allostatic changes. Selected lines such as crossed High Alcohol Preferring (cHAP) mice voluntarily drink to blood alcohol concentrations comparable to those attained in CIE and could be used to assess how alcohol affects these same endpoints without the confounds of involuntary vapor inhalation (Matson & Grahame, 2013). In three experiments, we assess how 2BC drinking in cHAP affects abstinence-related depressive- and anxiety-like behavior, operant responding for alcohol, and binge consumption using drinking-in-the-dark (DID). We hypothesized that cHAPs with home cage drinking experience would exhibit more depressive behavior after abstinence, increased responding for alcohol in the operant box, and increased DID intake. Of these, a drinking history increased DID intake in female cHAPs only and increased sucrose preference and intake following abstinence but had no effects on operant responding or NSFT latency and FST immobility following forced abstinence. These results are consistent with recent findings using slice electrophysiology showing tolerance to alcohol’s actions on the dorsolateral striatum following 2BC drinking in female, but not male cHAP mice (Rangel-Barajas et al., 2021). Overall these data suggest that cHAPs may require procedures allowing rapid intoxication such as DID, to demonstrate changes in alcohol’s rewarding effects.

Introduction

Many diagnostic criteria for alcohol use disorder (AUD) require repeated experience with alcohol over an extended period of time. For example, frequent drinking bouts allow for the development of chronic tolerance to ethanol, necessitating that the individual drink increasing amounts to achieve the same level of intoxication (Cappell, 1981). This can account for the observed escalation in drinking in patients with AUD, such as when they report “drinking more, or longer, than intended” (American Psychiatric Association, 2013). Individuals with AUD may also be so preoccupied with obtaining and consuming alcohol-containing drinks that these thoughts and behaviors interfere with their daily life and tasks; cues that have become associated with the availability of alcohol can also heighten such cravings (Fox et al., 2007). These associations, like chronic tolerance, require a history of stimulus pairings with alcohol to become cemented. Finally, though one can suffer acute withdrawal from a single alcohol binge, individuals with AUD who abstain from drinking may experience severe and life-threatening withdrawal symptoms such as hallucinations and seizures (Zador & Hall, 1997). This abstinence-induced withdrawal can extend to more emotional symptomatology as well; current frameworks of how AUD alters emotional homeostasis suggests that repeated ethanol exposure can effectively downshift one’s baseline affective state. This allostatic drift model suggests that recurrent drinking is driven by negative reinforcement such that ethanol consumption removes the aversive negative emotional state (Koob & Le Moal, 2001). Regardless of the theoretical framework, it is clear that a history with ethanol consumption is required for an AUD diagnosis because this pattern of ethanol exposure alters metabolic, motivational, and affective homeostasis, all potentiating disordered drinking.

In preclinical rodent studies, a pharmacologically relevant alcohol history is often achieved using the chronic intermittent ethanol (CIE) vapor paradigm whereby animals are exposed to ethanol vapor for 16 hours per day over 4 days or longer. During vapor exposure, mice and rats can reach intoxicating blood ethanol concentrations (BECs) between 175–225mg/dL (Becker & Lopez, 2004). Compared to vapor naïve controls, vapor-exposed animals have been shown to exhibit higher levels of responding to self-administer ethanol (Lopez & Becker, 2014; Vendruscolo & Roberts, 2014). These animals have also been shown to develop chronic tolerance to ethanol’s aversive effects (Lopez et al., 2012), neuroinhibitory effects (Nimitvilai et al., 2016), and ataxic effects (Daut et al., 2016). Additionally, CIE animals exhibit signs of acute withdrawal analogous to human AUD patients who quit drinking abruptly as measured by handling induced convulsions (HICs) (Becker & Hale, 1993). Following protracted abstinence from ethanol vapor, CIE animals also express a depressive and/or anxious phenotype as measured via behavioral assays like the Porsolt forced swim test (Walker et al., 2010) and elevated zero maze (Kliethermes et al., 2006), respectively. These data have been used to solidify CIE as a reliable method to induce an ethanol history in preclinical rodent models, but some critics argue against its use, mainly citing the method’s poor face validity due to its involuntary administration route (Spanagel, 2017). Additionally, while CIE BECs are high and relevant to studying the effects of intoxicating amounts of ethanol, they can be difficult to regulate and may vary considerably across individuals and strains (Lopez et al., 2011). Some CIE procedures also include daily injections of pyrazole, an alcohol dehydrogenase inhibitor, to promote stable BECs (Lopez et al., 2012). Finally, repeated exposure to forced high ethanol vapor can be stressful for animals, confounding the interpretation of these results when compared against an unstressed, vapor naïve control group, especially when endpoints include measures of affective disturbance.

To avoid the face validity and potential stress confounds associated with CIE, many experimenters have utilized the high-drinking C57BL/6J (B6) inbred strain along with voluntary home-cage or drinking-in-the-dark (DID) procedures to allow animals to voluntarily administer alcohol without the stressful confounds of vapor. While the consistency of history effects of B6 mice has been mixed using intermittent and continuous 2BC (Bloch et al., 2020; Bloch et al., 2022), several studies have demonstrated that a prolonged drinking period followed by a forced abstinence period results in increases in depressive and anxious behavioral correlates of affective disturbance when compared with water-drinking controls (Vranjkovic et al., 2018; Stevenson et al., 2009; Pang et al., 2013). Additionally, B6 mice with access to voluntary ethanol consumption have exhibited tolerance to the ataxic effects of ethanol (Cronise et al., 2005; Linsenbardt et al., 2011). Dependence phenotypes which include the increase in operant self-administration of ethanol and increased HICs seen following CIE are not commonly studied or observed following voluntary home-cage drinking in B6 mice (Phillips et al, 1994). This may stem from the relatively modest BECs these mice encounter during two-bottle choice (2BC; about 50–60 mg/dL; Matson and Grahame, 2013), as well from the fact that this strain, like most high-drinking populations, shows relatively low dependence vulnerability compared to other murine populations (Metten et al., 1998). Indeed, most of the literature citing changes in these metrics use CIE to induce the history of ethanol exposure (Chu et al, 2007; Lopez et al, 2008). This leaves a gap in the literature whereby the effects of a high-dose alcohol drinking history on self-administration in high-drinking animals are underexplored. Such studies may be useful for understanding the role that changes in alcohol seeking and alcohol reward play in the development of alcoholism. In addition, we believe that preclinical research on alcoholism should not rely almost entirely on any single mouse model. B6 mice, like any inbred strain, are a genetic N of 1; such homogeneity may reduce reproducibility (Bodden et al., 2019) without reducing within- or between-experiment variance (Tuttle et al., 2018), and their performance on a slew of assays has been highly capricious depending on small tweaks to environmental variables (Richter et al., 2011). Using animals that voluntarily drink intoxicating amounts of ethanol but that are not inbred to answer questions about the effects of a drinking history could be beneficial to fill this gap.

Selectively bred lines of mice like the High Alcohol Preferring mice (HAP) and the crossed High Alcohol Preferring mice (cHAP) may be more representative of individuals diagnosed with AUD in that they have a family history of high drinking, and were generated from heterogeneous progenitors, resulting in heterozygosity at loci unrelated to the high drinking phenotype. These lines show elevated two-bottle choice (2BC) preference and intake of ethanol. (Matson & Grahame, 2013; Grahame et al., 1999; Oberlin et al., 2011) On average as a population, cHAP mice reach CIE-level BECs (> 175mg/dL) in 2BC procedures, especially during their peak wakefulness hours between 6 – 10 hours into the dark part of their light cycle (Matson et al., 2013). Additionally, the cHAPs develop metabolic tolerance (Matson et al., 2013) and ataxic behavioral tolerance (Matson et al., 2014) following free-choice drinking. These animals, like B6, also respond for ethanol after instrumental conditioning in operant self-administration (O’Tousa et al., 2015), as well as showing intoxicating, binge levels of alcohol intake during a 2h Drinking in the Dark (DID) session (Ardinger et al., 2021). Similarly to B6 mice, HAPs show a minimal HIC responses following CIE-exposure, consistent with the repeatedly observed relationship between high volitional intake and less severe withdrawal effects (Metten et al, 1998). Although this lessens their utility for the study of alcohol withdrawal, their lack of nonspecific impairment would necessarily occur during severe withdrawal, such as seizures, also means that they may serve as a good candidate for observing the effect of a drinking history on affective, reward-related, and tolerance-related endpoints.

It should be briefly noted that in principle, changes in operant responding for ethanol could be driven by sensitization to alcohol’s stimulating effects. Like low withdrawal susceptibility, locomotor sensitization to alcohol tends to be positively correlated with propensity to drink (Grahame et al., 2000; Melón & Boehm, 2011), and has been observed in humans with an AUD (King et al., 2022), so it is entirely possible that altered alcohol self-administration following a history of drinking in cHAP mice may be caused by sensitization to some of ethanol’s actions.

In the following set of experiments, we utilized cHAP mice to explore the effect of a continuous, voluntary, high dose drinking history on outcomes that had previously been measured following either CIE vapor inhalation and/or the relatively modest exposure seen with voluntary consumption in B6 mice. Our main interests were affective disturbance and reward-related tolerance or sensitization. First, we examined the effect of a 5-week drinking history and subsequent forced abstinence periods of varying lengths on affective disturbance assays that have emerged from B6 mice voluntarily consuming alcohol (Holleran & Winder, 2016). Using drinking parameters validated to cause affective changes following abstinence (Vranjkovic et al., 2018), we hypothesized that voluntary consumption would induce symptoms of abstinence-related depressive-like behaviors after one or both of the abstinence lengths. Second, we explored the effect of a 5-week drinking history on subsequent instrumental behavior, analogous to studies showing changes in these endpoints following chronic or intermittent ethanol access (Somkuwar et al., 2016; Chu et al., 2007). We hypothesized that a drinking history would increase operant oral self-administration of alcohol as well as responding in extinction; the former might be interpreted as tolerance or sensitization to ethanol’s acute rewarding effects, while the latter could be construed as a measure of craving or alcohol seeking that would be uncomplicated by intoxication during the post-dependence testing. Finally we explored the effects of a 14-day home cage ethanol drinking history on subsequent DID drinking, hypothesizing that chronic alcohol consumption would increase intake during binge ethanol access.

Methods

Subjects

Each of the experiments utilized independent groups of ethanol-naïve cHAP mice: 50 (24 F, 26 M) mice from the 46^th generation of selection and 36 (23 F, 13 M) mice from the 52^nd generation of selection for Experiment 1, two balanced cohorts of 33 (16 F, 17 M) and 36 (18 F, 18 M) mice from the 47^th (cohort 1) and 48^th generations (cohort 2) for Experiment 2, and 32 (16 F, 16 M) mice from the 49^th generation for Experiment 3. All mice were born in the IUPUI Animal Care Facilities. One week prior to any experimental manipulation, all animals were single-housed in a 12:12 reverse light-dark cycle at 65–87 days of age. When in their home cages, animals had ad libitum access to chow and water, except where indicated below. Mice were provided with pine bedding and cages were changed biweekly by the experimenter. Mice were counterbalanced by sex and family into drinking history groups and abstinence length groups for all experiments, as well as into squads of 12 for Experiment 2 (see Operant Apparatus). All work was conducted at an AAALAC-approved facility adhering to the National Research Council’s Guide for the Care and Use of Laboratory Animals.

Ethanol Solutions

For 2BC and operant training and testing, ethanol was prepared by diluting 200 proof ethanol (Pharmco Inc.; Brookfield, CT) to 10% v/v in tap water. For the DID procedure, a 20% v/v solution was used from the same 200 proof stock. This 20% ethanol solution was stored in sealed reservoirs connected to the volumetric drinking monitor (VDM) system. (see Drinking in the Dark).

Free-Choice Drinking

Mice were weighed weekly for accurate gram-per-kilogram intake calculations. One week after single-housing, home-cage water bottles were replaced with a 25mL and a 50mL graduated cylinder readable to ± 0.5mL. For ethanol history animals, the larger cylinder contained 10% (v/v) ethanol while the other was filled with tap water. For water history animals, both cylinders contained tap water. 2BC was continuous, as in bottles remained on the cage for the entire drinking period (35 or 14 days) Three times a week, intakes were measured on the home cage to prevent leaks and spills and cylinder positions were swapped to control for side preference. In Experiment 1, animals had 35 days of free-choice drinking before an abstinence period of either 1 or 3 weeks; during abstinence, all cylinders were replaced with home-cage water bottles. In Experiment 2, once an animal reached criteria in FR8 training (see Operant Training) they transitioned to 2BC, and animals had 35 days of free-choice drinking history before operant testing. Experiment 3 included 0- and 14-day drinking histories before DID. Here we opted for a 14-day rather than 35-day drinking history based on a preliminary finding in our lab that 14 and 35-days of 2BC drinking yielded equivalent intakes in subsequent DID drinking in a different selectively-bred mouse line. Thus, we opted for the shorter history here, which proved effective (see results).

Affective Disturbance Assays (Experiment 1)

Following one or three weeks of abstinence from free-choice drinking, animals underwent a home cage sucrose preference test (SPT). For 24 hours, each subjects’ water bottle was replaced with two 10mL graduated cylinders, one containing a 1% sucrose solution and the other containing tap water. Fluid levels were measured at placement, at the 12-hour point, and at the 24-hour point before the cylinders were replaced with the home cage water bottle. At the 12-hour reading, cylinders were side-swapped to control for positional preference. This procedure was adapted from Xu et al. (2022) who found a decrease in sucrose preference after 8 weeks of HC drinking, though the 24-h deprivation period was not utilized here.

The novelty-suppressed feeding test (NSFT) protocol began 24h after sucrose removal. At this time, all food was removed from the home cages for a 48h food deprivation period with a free-feeding window at 23–25h. Mice were brought into the testing room for 1h before NSFT to acclimate. Testing occurred during the dark part of the animals’ cycle; thus red illumination was used. One at a time, animals were placed in the same corner of a large white box (40cm × 40cm × 38cm) containing 2cm of standard pine bedding. One pellet of standard chow was placed in the center of the box. Mice were allowed to explore for up to 5 minutes maximum. Mice were recorded and tracked by Any Maze software, and the latency to take the first bite of the center food pellet was recorded for each animal by the experimenter. Once this occurred, the animal was removed from the box and returned to the home cage, and the recording was stopped. Animals were then given fresh food in the home cage, completing their deprivation period.

The forced swim test (FST) occurred 24h after NSFT completion. Four 2-liter beakers were filled halfway with water, placed on a rack across from a video camera, and allowed to come to room temperature (23–25 °C) for one hour. Mice were brought into the testing room for 1h to acclimate. At testing, four mice at a time were placed in the water in separate beakers and filmed for six minutes. After the allotted time, animals were placed back in their home cages with a paper towel for drying, which were removed before the end of the test day. Immobility time in the last four minutes of the swim was scored by a blind observer. FST and NSFT procedures were adapted from previous studies showing that forced abstinence from HC drinking induces affective disturbance in B6 mice (Holleran et al., 2016; Vranjkovic et al., 2018). Figure 1A illustrates the timeline for Experiment 1.

Figure 1.

Experimental Timelines for (A) Experiment 1, (B) Experiment 2, and (C) Experiment 3. Figure created with BioRender.com

Operant Apparatus

For Experiment 2, twelve operant chambers (Med-Associates) with dimensions of 21.6 × 19.7 × 12.7 cm were used simultaneously for each squad. Each chamber was situated in its own light- and sound-attenuating box with a fan positioned in the center of the right-facing wall relative to the front door of the box to promote air flow during training and testing. A retractable sipper tube attached to a 10mL graduated pipette could provide a 10% ethanol solution to serve as a reinforcer; readability to ± 0.05 ml allowed accurate measurement of alcohol intake following these sessions. Animals drink directly from the sipper when it is presented; thus, direct intake is measured rather than rewards granted. On either side of the sipper and on the same wall, two non-retractable levers (active versus inactive) were available to manipulate. Lever presses and sipper deliveries (rewards) were tracked by Med-PC IV software (Med-Associates). For training and testing, no house lights or lever lights were utilized. “Yesterday’s News” bedding covered the bottom of the compartment under the grid floor of each chamber and was changed weekly. Operant chambers were cleaned with 70% ethanol before each squad each day.

Water Deprivation (Experiment 2)

In Experiment 2, animals were fluid deprived for 22 hours before training to encourage drinking behavior in early operant training. Water bottles were returned to the home cage for two hours before being removed again before the following session. Water deprivation occurred only for the fixed time 120s (FT120s) sessions and the first fixed ratio 1 (FR1) session for each animal, described below; otherwise, water was available ad libitum in the home cage at all times from the second FR1 session onward, but not during operant sessions.

Operant Training (Experiment 2)

In Experiment 2, operant training began with one 30-minute FT120s magazine training session where all active lever presses were reinforced with 20s of access to the sipper containing 10% ethanol. The sipper was also presented every 120 seconds regardless of the subject’s input. Animals that did not consume at least 0.2mL of fluid received additional FT120s sessions until they reached criterion.

Animals then moved to FR1 training sessions for 60 minutes where each lever press on the active lever was reinforced by the presentation of the ethanol sipper for 5 seconds while inactive lever presses resulted in nothing. Levers were not retracted during sipper access, but after the sipper retracted, response-contingent reinforcement was available immediately without a timeout. Lever presses occurring while the sipper was extended were not recorded. Mice had a minimum of three FR1 training sessions and were moved on to FR3 when they consumed at least 0.2mL on average over three sessions.

During the 60-minute FR3 sessions, mice had 12s of sipper access when the active lever was pressed 3 times, regardless of the length of time between presses or number of inactive lever presses between active lever presses. When animals consumed an average of at least 0.2mL over at least three sessions, they advanced to FR6 sessions. FR6 and FR8 sessions were the same length as FR1 and FR3 sessions with the same completion criteria. Variability of the three most recent sessions of FR8 was required to be lower than 20% for each subject to be included moving forward.

If an animal became stuck in FR1 or FR3 training due to low intake, water deprivation as described above was used to encourage drinking in the following training session. If an animal was held up in FR6 or FR8 training, those specific subjects were run on back-to-back sessions that same day until they reached criterion, and then training continued on the next day without additional fluid deprivation. Retraining methods failed to produce lasting patterns of drinking in two male animals in cohort 2, and these animals were excluded from the study.

Reinforced and Extinction Testing (Experiment 2)

For Experiment 2, three hours before testing, 2BC cylinders were removed from the home cage and replaced with standard glass water bottles. Mice were again split into two counterbalanced groups based on sex, family, and drinking history (W or E). One group received a reinforced FR8 session on the test day, while the other group received an extinction test in the same operant box as training. Reinforced sessions were 2 hours in length while extinction sessions were 1 hour. Longer reinforced sessions were used here with the hope of increasing the likelihood of detecting an alcohol history effect on persistence of responding during longer sessions, which may yield higher, more intoxicating BECs. During extinction testing, 8 lever presses on the active lever caused an empty sipper to descend into the operant chamber. During both reinforced and extinction testing, active and inactive lever presses were recorded, and during reinforced sessions, intake was measured by reading the volumetric sipper tube at the end of the session. Figure 1B illustrates the timeline for Experiment 2

Drinking-in-the-Dark (Experiment 3)

For Experiment 3, three days prior to the first DID session, animals were transported to the testing room in their home cages to acclimate. On the first day of DID testing, alcohol cylinders were removed at lights-off at 9AM (3 hours prior to DID to assure that animals were not intoxicated at the time of testing). At 12PM, water cylinders were removed and replaced with a volumetric drinking monitor sensor (VDM; Columbus Instruments Inc., Columbus, Ohio, United States) connected to a vessel of 20% v/v ethanol. Animals had access to this sipper for 2h and intake was recorded via the VDM software in 1-minute bins. Figure 1C illustrates the timeline for Experiment 3.

Statistics

Free-choice ethanol intake was analyzed using repeated measures 3-way mixed-method ANOVAs. To check for escalation of drinking, each animal’s first two and last two intake measures were averaged together, respectively, and compared via a paired t-test. Sucrose preference, novelty suppressed feeding latency and forced swim immobility time were analyzed via 3-way ANOVAs and collapsed on sex if there was no main effect. Correct lever presses, incorrect lever presses, and intake in the operant chamber or DID test were analyzed via 2-way ANOVAs (group-by-sex). If main effects or interactions were detected, Tukey’s post-hoc multiple comparisons were utilized to reveal differences between groups split by sex. Outliers were detected using Grubb’s test and removed if detected for drinking data only (sucrose drinking, ethanol drinking) since leaks may be a source of measurement error. Outliers in other behavioral variables have not been removed. In Experiment 3 only, frontloading during DID was quantified by assessing the percentage of total, 2-h intake consumed during the first 15 min (Ardinger et al., 2021). We further analyzed the cumulative intake patterns of each mouse by using 1-min binned data from the VDMs to assess whether an animal frontloaded or not based on criteria outlined in a recent Ardinger et al. (2022) review purporting frontloading as a behavioral correlate of motivation to drink alcohol. For this, we utilized a freely available Matlab script which applies a parcs algorithm to identify a change point and then classifies the animal as a “frontloader” or “non-frontloader/inconclusive” based on the temporal location of the change point in the session and the slope of the pre-change point intake. A chi-square test was then applied to detect any differences in the distribution of frontloaders between water history and alcohol history mice.

Results

Free-Choice Drinking

Initial and final home cage alcohol intakes for all experiments are shown in Table 1.

Table 1.

Initial and final intakes and preferences across both sexes in each experiment. Initial and final readings are the means of the first two and the last two readings, respectively, for each animal. Intake is the top line in each cell, while percent preference is the bottom line in each cell. Data shown are mean ± SEM. If significant escalation was detected via paired t-test, the Final reading cells are marked as follows:

	Female		Male
	Initial Intake (g/kg/24h) & Preference	Final Intake (g/kg/24h) & Preference	Initial Intake (g/kg/24h) & Preference	Final Intake (g/kg/24h) & Preference
Expt. 1 (35 days)	26.49 ± 0.91 82.25% ± 1.86	29.11 ± 0.99* 94.83% ± 0.81****	22.71 ± 1.06 76.44% ± 3.06	24.49 ± 1.57 91.25% ± 2.38**
Expt. 2 (35 days)	23.82 ± 0.94 83.25% ± 2.79	27.12 ± 0.48** 94.19% ± 0.99**	21.93 ± 0.99 82.19% ± 2.67	22.75 ± 0.65 94.13% ± 0.67***
Expt. 3 (14 days)	24.83 ± 1.13 85.38% ± 3.19	28.81 ± 0.66* 95.13% ± .57*	20.74 ± 0.91 76.34% ± 4.58	27.17 ± 1.10** 92.91% ± 1.47**

^*p<0.05,

^**p<0.01,

^***p<0.001,

^****p<0.0001.

Experiment 1.

A three-way (group-by-sex-by-reading) mixed-methods ANOVA of free-choice ethanol intake in the home cage revealed a difference across reading days (F[14, 525] = 6.01, p < 0.0001) indicating that drinking fluctuated significantly across time. Female cHAPs drank more than males (F[1, 39] = 9.27, p = 0.0042) and there was no 3-way interaction (p > 0.05). Abstinence group was included as a factor to check for intake differences between the abstinence groups prior to the introduction of abstinence. There was a main effect of abstinence group (F[1, 39]) = 7.30, p = 0.01). An unpaired t-test collapsed across sex between the 1-week and 3-week abstinence groups analyzing averages of the last two days of drinking in the home cage revealed that 1-week abstinence animals had higher intakes prior to the start of abstinence (t[40] = 2.752, p = 0.009). Animals in experiment 1 escalated drinking significantly according to a two-tailed, paired t-test comparing initial and final readings (t[35] = 4.094, p = 0.0002).

Experiment 2.

A three-way (cohort-by-sex-by-reading) mixed-methods ANOVA of ethanol intake revealed main effects of reading days (F[14, 398] = 5.85, p < 0.0001) and sex, with female cHAPs drinking more than males (F[1, 29] = 8.92, p = 0.006) and no 3-way interaction (p > 0.05). Cohort was included as a factor to ensure no intake differences between cohorts, which was the case (p > 0.05), though there was a significant cohort-by-reading interaction (F[14, 398] = 2.24, p = 0.006). An unpaired t-test collapsed across sex between both cohorts analyzing averages of the last two days of drinking revealed no differences in intake before operant testing (p > 0.05). Animals in experiment 2 did escalate according to a paired t-test comparing first and last readings, collapsed across sex (t[31] = 2.28, p = 0.003).

Experiment 3.

A two-way mixed-methods ANOVA revealed a main effect of reading (F[6, 83] = 19.92, p < 0.0001) and sex (F[1, 14] = 5.71, p = 0.03) with female cHAPs outdrinking males. There was no interaction (p > 0.05). A paired t-test collapsed on sex illustrates significant escalation (t[15] = 5.38, p < 0.0001).

Affective Disturbance Assays (Experiment 1)

Overall, the SPT was the only affective test to show an effect of history, though in the opposite direction of our hypothesis, as animals with an alcohol history drank more sucrose than animals without that history. A three-way ANOVA on preference scores (abstinence length-by-history-by-sex) revealed only a main effect of history (F[1, 77] = 4.93, p = 0.03) and no interactions. Additionally, the same analysis on sucrose intake (g/kg) revealed a main effect of history (F[1, 76] = 10.38, p = 0.002) and of abstinence length (F[1, 76] = 24.47, p = 0.0002) and no interactions. A female in the 1-week water group (11.74g/kg) was removed here using Grubb’s test. Thus, a 5-week history of alcohol in the home cage increased subsequent preference and intake of sucrose in 1-week abstinence animals. Figure 2A sucrose preference collapsed on sex, while figure 2B shows gram-per-kilogram sucrose intake.

Figure 2.

Experiment 1, Affective Disturbance Assays. All graphs are displayed as mean ± SEM. A: Sucrose preference collapsed across sex. Sucrose preference was higher in ethanol history animals regardless of history length (p = 0.03) and this effect was significant in the 1-week group via pairwise comparison (Tukey’s, p = 0.01). B: Sucrose intake collapsed across sex. Animals abstinent 1-week drank more 1% sucrose then 3-week animals (p = 0.0002) and alcohol history animals drank more than water animals (p = 0.002) and this effect was significant in the 1-week group via pairwise comparison (Tukey’s, p = 0.003). C: Novelty suppressed feeding latency collapsed across sex. There were no differences between history or abstinence groups. D: Forced swim immobility time split by sex and drinking history across abstinence lengths. Males were more immobile than females (p = 0.0004) and 3-week animals were more immobile than 1-week animals (p = 0.04) with an interaction on sex and abstinence length (p = 0.014). Drinking history, however, had no effect on this behavior. *p < 0.05, **p < 0.01 ***p < 0.001

In the NSFT, a three-way ANOVA on latency to first feeding (abstinence length-by-history-by-sex) revealed no main effects or interactions (p’s > 0.05). Figure 2B illustrates NSF latency collapsed on sex.

In the FST, a three-way ANOVA on immobility time (abstinence length-by-history-by-sex) revealed only a main effect of sex (F[1,73] = 13.62, p = 0.0004) and abstinence length (F[1, 73] = 4.26, p = 0.04) and an interaction between sex and abstinence length (F[1, 73] = 6.37, p = 0.014). Importantly, there was no main effect of drinking history here, and no interactions between drinking history and sex or abstinence (p > 0.05). Figure 2C shows FST immobility.

Reinforced and Extinction Testing (Experiment 2)

Animals proceeded from FT120s to FR8 sessions as described above. Across both balanced cohorts, the mean number of sessions from the beginning of training to the final FR8 criterion was 12.67 sessions. Average active and inactive lever presses and ethanol intake from the last three FR8 sessions of training are displayed in Table 2. These data were analyzed with two-way ANOVAs with factors of sex and history to ensure no differences between history groups prior to the two-week HC drinking phase. During FR8 training, females responded more on the active lever (F[1, 63] = 6.27, p = 0.015) and drank more than males (F[1, 63] = 22.39, p < 0.0001). There were no such sex effects in inactive lever pressing and no history main effects or interactions (p’s > 0.05).

Table 2.

Active and inactive lever press responses and 10% ethanol intake averaged across the last three days of FR8 training.

	Female			Male
	Active*	Inactive	Intake**** (g/kg)	Active*	Inactive	Intake**** (g/kg)
Water	111.58 ± 21.51	33.91 ± 4.74	2.73 ± 0.12	67.48 ± 7.09	42.42 ± 7.08	1.57 ± 0.20
Ethanol	105.87 ± 17.55	41.16 ± 3.66	2.60 ± 0.22	72.91 ± 8.93	42.12 ± 5.50	2.01 ± 0.20

Females responded more on the active lever (*p < 0.05) and drank more (****p < 0.0001) than males. Data shown are mean ± SEM.

For analysis of test day data, separate two-way ANOVAs were run for the reinforced group and the extinction group, each using factors of sex and history. For animals that were reinforced (Fig. 3), females responded more (F[1, 30] = 5.09, p = 0.03) and drank more (F[1, 30] = 12.57, p = 0.001) than males. There were no effects of or interactions with drinking history (p’s > 0.05). There were also no differences in inactive lever pressing and no interactions (p’s > 0.05)

Figure 3.

Experiment 2, Reinforced Operant Responding. All graphs are displayed as mean ± SEM. A: Active lever presses across sex and history group. Females pressed more on the active lever than males (p = 0.03). B: Inactive lever presses across sex and history group. C: Ethanol intake across sex and history group. Females drank more than males during the reinforced test. *p < 0.05, **p < 0.01

For animals tested in extinction, active lever presses showed neither effects of an alcohol drinking history, differences between the sexes, nor any interactions (p > 0.05). Unexpectedly, however, there was a main effect of history group in inactive lever pressing (F[1, 28] = 4.89, p = 0.04) and a significant sex-by-history interaction (F[1, 28] = 5.77, p = 0.02). Tukey’s post-hoc multiple comparisons indicated that female ethanol history animals pressed the inactive lever significantly more frequently than their female water history counterparts (p = 0.02), a finding that was not observed among male cHAPs. Figure 4 illustrates these extinction results.

Figure 4.

Experiment 2, Extinction Responding. All graphs are displayed as mean ± SEM. A: Active lever presses across sex and history group in extinction. B: Inactive lever presses across sex and history group in extinction. A main effect of history (p = 0.04) and a significant interaction (p = 0.02) are driven by the fact that ethanol history females responded more frequently on the inactive lever than water history females (Tukey’s, p = 0.02). *p < 0.05

DID (Experiment 3)

DID intake and percent of intake in the first 15 minutes were analyzed using two-way ANOVAs with factors of sex and history. There was a main effect of history on intake during the DID session (F[1, 28] = 11.47, p = 0.0021) and a significant interaction (F[1, 28] = 4.88, p = 0.04); further analysis with Tukey’s post-hoc comparisons revealed that female cHAPs with an ethanol history drank more in DID than their ethanol naïve counterparts (p = 0.003) (Figure 5A), while males showed no effect of a drinking history (p > 0.05). Regardless of sex, cHAPs with a water history frontloaded more of their total consumption than alcohol history animals (F[1, 28] = 9.48, p = 0.005) (Figure 5B), indicating that only alcohol history animals continued to drink after the first 15 minutes of the DID session. This is confirmed by analyzing the total intake in g/kg in the first 15 minutes (Figure 5C) whereby there are no differences between sex or history group and no interaction (p’s > 0.05).

Figure 5.

Experiment 3, DID Intake and Frontloading. All graphs are displayed as mean ± SEM. A: Total intake across sex and history groups during the 2hr DID session. Female cHAPs with an alcohol history drank more than female cHAPs with a water history (p = 0.003). B: Percent of intake in the first 15 minutes of the DID session across sex and history groups. Male and female cHAPs frontload significantly more without an alcohol history (p = 0.005). C: Intake in the first 15 minutes across sex and history groups. No differences. D: Intake from the remainder of the DID session, after the first 15 minutes. Animals with a history of ethanol in the home cage drank more after the first 15 minutes than water history animals (p = 0.0016) and pairwise comparisons reveal a significant effect in females (p = 0.03). *p < 0.05 **p < 0.01

To further understand how a drinking history may modulate drinking patterns, 1-min binned VDM data from the DID session was processed via the algorithm available in a recent Ardinger et al. review (2022). Animals were classified as frontloaders vs. non-frontloaders and a chi-square test was applied across between these outcomes and drinking history. No differences were detected here (p > 0.05), though this is likely due to the fact that out of 29 total animals analyzed, 24 significantly frontloaded according to the algorithm. This is congruent with previous data on drinking patterns in cHAP mice who tend to frontload heavily upon their first binge opportunity (Ardinger et al., 2021). Since most of the animals frontload regardless of history, we plotted the remaining intake in Figure 5D, that is, the intake after the first 15 minutes of the session, split by sex and history. Here we find a main effect of history (F[1, 27] = 18.73; p = 0.0016) and no interaction, demonstrating that ethanol history animals continue to drink significantly more than water animals for the remainder of the session.

Discussion

Selectively-bred cHAP mice in these experiments drank significant quantities of ethanol in the home cage during 2BC, similar to free-choice intakes observed in Matson et al. (2013) that resulted in average blood ethanol concentrations exceeding 250 mg/dl. These intakes have been sufficient to cause behavioral (Matson et al, 2014) and metabolic (Matson et al., 2013) tolerance, and compulsive alcohol consumption as measured by persistent intake in the presence of the bitter tastant, quinine (Houck et al., 2019). Thus, we were able to address the main questions of this series of studies: does a history of very high, free-choice drinking in an outbred animal model lead to abstinence-related affective disturbance, sensitization or tolerance to alcohol reinforcement, changes in craving as measured by extinction, or changes in intake as measured by the DID procedure? We observed that drinking histories similar to those causing behavior changes in prior experiments did not cause depressive-like behaviors in the FST or NSFT following abstinence or on operant oral self-administration but had sex-specific effects in the DID binge-drinking model. There was, however, an interesting history effect in the SPT, whereby ethanol history animals exhibited increased sucrose preference, which is the opposite phenotype observed in animals considered affectively disturbed.

Although intakes in these mice are far higher than those that result in depression-like changes in B6 mice (Holleran et al. 2016), the results from Experiment 1 indicate that a 5-week history of home cage drinking has incongruent effects on sucrose preference, novelty suppressed feeding latency, or immobility in the forced swim test following an abstinence period in cHAP mice. In the SPT, access to alcohol in the home cage actually increased both sucrose preference and intake, a finding that suggests cHAP mice exposed to alcohol have an increased hedonic response to sucrose. Adult B6 mice have also shown increased preference, however, after a history of drinking (Lee et al., 2017). To control for a possible caloric deficit that alcohol history animals may be experiencing, we suggest repeating this test with saccharin instead of sucrose. Despite a sex effect indicating that males were more immobile in the FST than females and a history effect whereby 3-week abstinent animals were more immobile than 1-week animals, there was no effect of home-cage drinking history on forced swim immobility. These findings contrast the withdrawal effects that are observed in B6 mice after an equivalent home cage drinking period or following CIE exposure (Stevenson et al, 2009, Jury et al, 2018). However, these results in cHAPs may be congruent with CIE exposure’s minimal effect on withdrawal induced HICs in the related HAP line (Lopez et al., 2011). Thus, it may be that high drinking selected lines like the cHAPs and HAPs are not as susceptible to commonly observed endpoints of alcohol withdrawal, including depression-like behaviors, or that corresponding 2BC drinking histories or CIE exposure levels in B6 mice must be elongated or intensified to have similar effects in HAPs and cHAPs. Indeed, Xu et al. (2021) have shown that female cHAPs display an anxiety-like phenotype in the elevated plus maze assay after a 7-month home cage drinking history, although that was also associated with substantial thiamine deficiency and neurotoxicity. Additionally, the fact that a high-drinking line like the cHAPs has low susceptibility to ethanol withdrawal aligns with previous work that highlights the genetically based relationship between high drinking and low withdrawal sensitivity (Metten et al., 1998). One caveat of our study that needs mentioning are the differences between the protocols used here compared to those commonly performed to study affective disturbance. Some labs perform the SPT over an 8-day period whereby the first four days involve an acclimation period, the next two days involve baseline measurements, and the final two days are a 24-h food and water deprivation period and the actual test (Liu et al, 2018). Recall that we were interested in observing abstinence effects at two timepoints: 1 week and 3 weeks post-drinking. Because 48-h food deprivation was also a necessary component of the NSFT which delayed FST testing to 6 days following SPT, ultimately lengthening abstinence at the time of the forced swim, we opted to use a simple home-cage version of the SPT which just involves replacing home cage water bottles with sucrose and water tubes without deprivation. Additionally, other groups perform the NSFT or FST in the light part of the cycle, introducing a baseline level of anxiety that may not be present in the dark (Vranjkovic et al, 2018; Sung et al, 2022). Indeed, recent research has shown that dark-cycle application of the NSFT increases hyperphagia and light-cycle application induces hypophagia in B6 mice, regardless of sex (Francois et al., 2022). Importantly, this study did not assess alcohol’s effect on NSFT latency in either day or night condition. Since the other experiments (operant testing and DID) in this battery of studies needed to be done in the dark part of the cycle, we opted for light cycle coherence between these cHAP experiments. Thus, these adjustments may account for the lack of history or abstinence effects, and future exploration of depressive or anxiety-like phenotypes in cHAPs could isolate specific assays and/or run them under standard lighting procedures.

In Experiment 2, a 5-week history of home cage 2BC in cHAPs did not have an effect on active lever presses in reinforced or extinction testing. Corresponding to their higher 2BC drinking, females pressed on the active lever more than males and drank more in the reinforced test, suggesting that female cHAPs are more motivated to respond for and drink ethanol than males, regardless of home cage drinking history. The lack of alcohol history effects observed here contrast with studies that use CIE to induce motivational differences between treatment groups; CIE-exposed animals respond more in the operant chamber for alcohol during withdrawal periods than alcohol naïve animals (Lopez & Becker, 2014; Vendruscolo & Roberts, 2014). Mice in the current study drank to levels comparable to those seen previously in the home cage, and thus should have reached CIE-relevant BECs, or higher, averaging over 250 mg/dl (Matson et al., 2013). It may seem then that the 5-week drinking history did not impart any sort of motivational tolerance on the cHAPs here, though some aspect of the experimental design may have caused us to miss such an effect. Firstly, ethanol was used as the reinforcer throughout training so as not to cause any positive or negative contrast effects when the reinforcer was presented during the test. Thus, water history animals were not completely alcohol naïve, having had at least eight 1-h operant oral sessions during which they consumed substantial doses of alcohol. Secondly, water deprivation prior to the FT120s session caused high drinking in all animals, possibly inducing aversive effects which may have lengthened the training period, lengthening the water history’s exposure to ethanol in training. The goal of this first session is not to produce heavy intoxication, but rather to familiarize the animals with the layout of the operant box. Using a water reinforcer here would have prevented any aversive association with alcohol and minimized the water group’s alcohol exposure history. Finally total intakes in 1-h training (average 1.95g/kg) and 2-h reinforced testing (average 1.99g/kg) are not as high as seen in DID where average cHAP intakes can reach 3–4g/kg in 2-h (Ardinger et al., 2021). The cHAP mice also tend to frontload in DID (as seen in Figure 5), skewing their intake toward the beginning of the session. This pattern of drinking may be indicative of the drive to reach intoxicating BECs quickly (Ardinger et al., 2022) and if animals cannot physically reach that level of intoxication in the operant box because by necessity, access to alcohol is intermittent rather than continuous as in DID, then they may not encounter the kinds of intoxication levels at which a tolerance-like effect of an alcohol history could be observed.

In the extinction testing in Experiment 2, females with an alcohol history pressed more on the inactive lever than water drinking control females. Extinction often results in an increase in behavioral variability, as demonstrated by Neuringer et al (2001). Thus, when contingencies are broken, animals may increase responding on alternative manipulanda. The fact that this increase was only observed in female mice might be because, as previously mentioned, female cHAPs seek and take alcohol more than male cHAPs. Downshifts in reward magnitude, which are similar to extinction, also increase behavioral variability (Carlton, 1962). This downshift may have been greater for ethanol history females during the extinction test, resulting in higher inactive lever presses. Nevertheless, this is at best an indirect measure of motivation for ethanol in these animals, as there were again no differences in active lever responding between history groups on the active lever in either test condition.

Finally, in Experiment 3, we demonstrate that a history of 2BC drinking causes female cHAPs to increase the amount consumed on their first DID drinking session, an effect that could be interpreted as tolerance (or, potentially, sensitization) resulting from a 2-week drinking history. Interestingly, recent slice electrophysiology work with these same animals shows that females, but not males, acquire tolerance to the acute effects of ethanol on glutamatergic signaling in the dorsolateral striatum following a drinking history (Rangel-Barajas et al. 2021). This may imply that tolerance to alcohol’s actions on this structure may be involved in the sex-specific escalated DID drinking that we observed here. Consistent with this idea is that although alcohol history animals do significantly frontload as per the definition that >12.5% of total intake occurs within the first 15 minutes of the 2-h session, water history animals frontload significantly more regardless of sex. One way to view this difference in drinking pattern is that without alcohol tolerance, water history mice drink quickly during their first DID session and are unable to continue drinking due to acute intoxication. However, although mice with alcohol experience drink a similar amount during this period, they are able to continue drinking for a longer period due to either ataxic tolerance or to a desire to continue drinking because of tolerance to alcohol’s rewarding actions. Notably, cHAP mice acquire ataxic tolerance following as little as three days of free-choice alcohol consumption (Matson et al., 2014). Another possibility is that increased drinking is driven by metabolic tolerance, which we would need to explore in future studies that obtain BECs on the DID test day. The contrast between the results of Experiments 2 and 3 highlight what may actually be most reinforcing to the cHAP line: rapid intoxication, which may be precluded during operant oral self-administration session – or at least, the procedures used here, such as a relatively lean FR-8 reinforcement schedule. Overall, however, we cannot in these studies distinguish between a permissive tolerance (i.e., the ability to withstand ataxia/affective changes) or a tolerance (or sensitization) to ethanol’s attractive, rewarding effects as being key in driving escalated DID and 2BC intake.

In conclusion, a home cage alcohol history has differing effects on commonly measured outputs of dependence in cHAP mice. Following a 5-week alcohol drinking history and either 1 or 3 weeks of abstinence, cHAP mice actually show increased sucrose preference and intake, no evidence of novelty suppressed feeding, or increased immobility in the FST compared to water drinking controls. Abstinence length and sex did modulate immobility in the FST, however. The same 5-week drinking history did not have an effect on reinforced operant responding, though females responded more on the active lever and drank more alcohol, consistent with their higher 2BC intake. Female cHAPs with a drinking history did press more on the inactive lever in extinction, however, which could reflect the heightened effects of drinking history on reward downshift. Finally, female cHAPs with a 2-week drinking history drank more in a subsequent DID challenge than alcohol naïve females. The ability to detect this difference in DID and not in the operant experiments is likely due to the fact that cHAPs are motivated by rapid intoxication which is more easily attainable in DID than in an operant task.

Highlights.

• Up to 3 weeks of abstinence from 2BC drinking did not induce affective disturbance.

• Alcohol drinking experience did not change operant responding for alcohol.

• Alcohol drinking experience did not change extinction responding for alcohol.

• Alcohol experience increased intake in drinking-in-the-dark in female cHAPs only.

• During DID, alcohol experience caused more persistent binge-like behavior.