Dissecting operant alcohol self-administration using saccharin-fading procedure

Abstract

Background

Although alcohol use disorder is a complex human pathology, the use of animal models represents an opportunity to study some aspects of this pathology. One of the most used paradigms to study the voluntary alcohol consumption in rodents is operant self-administration (OSA).

Aims

In order to facilitate the performance of this paradigm, we aim to describe some critical steps of OSA under a saccharin-fading procedure.

Material & Methods

We used 40 male Wistar rats to study the process of acquiring the operant response through a saccharin-fading procedure under a fixed ratio (FR1) schedule of reinforcement. Next, we analyze the alcohol introduction and concentration increase, the effect of an alcohol deprivation, and the analogy between this paradigm with the Drinking in the Dark-Multiple Scheduled Access paradigm.

Results

During alcohol concentration increase, animals reduced their lever presses in accordance with the increase in alcohol concentration. On the contrary, the consumption measured in g·kg⁻¹ BW showed a great stability. The lever presses pattern within operant session changes with the introduction of different alcohol concentrations: at higher alcohol concentrations, animals tended to accumulate most of their presses in the initial period of the session.

Discussion

We show the utility of fading with low concentrations of saccharin and the evolution of the operant response through the different concentrations of alcohol.

Conclusion

Taken together, our results aimed to dissect the acquisition and maintenance of OSA behavior as well as other related variables, to facilitate the understanding and performance of this paradigm.

INTRODUCTION

Worldwide in 2016, around 2.3 billion people were current drinkers. Total alcohol per capita consumption in the world´s population over 15 years was 6.4 liters, with European countries being those that show the highest consumption. These data turned out in 3 million deaths (5.1% of all deaths) worldwide.¹ Alcohol use disorder is a complex human pathology, and it cannot be completely replicated under the conditions and restrictions of a laboratory. However, the use of animal models represents an opportunity to study several behavioral aspects of this pathology.² Animal models of alcohol consumption are used to provide valuable information for psychopharmacological treatments, as well as the description of the mechanisms associated with the consumption of this drug.^3,4 There are several methods to assess voluntary alcohol consumption in rodents, among which we find both models of moderate alcohol consumption, such as Operant Self-Administration (OSA), and models of excessive alcohol consumption.⁵

Operant Self-Administration is one of the most commonly used behavioral models to evaluate the positive reinforcement of self-administered alcohol.⁶ This paradigm has been shown to have great reliability and predictive validity in humans.⁷ In this paradigm, animals are trained to press a lever to obtain alcohol.⁸ It has been used not only for alcohol intake, but also to study the motivational characteristics of alcohol using of progressive ratio schedules.⁹ The most common schedule of reinforcement is the fixed ratio (FR) schedule, in which alcohol is delivered each time a preselected number of lever presses (ie, responses) have been completed. This kind of reinforcement is less affected by motivational factors than other schedules of reinforcement.^2,4 One of the main concerns presented by this paradigm is the time it takes for animals to acquire and maintain a stable operant response. This is due, at least in part, to the fact that a standard protocol to promote consumption through this paradigm requires a previous saccharin or sucrose fading for the acquisition of OSA.¹⁰ However, several studies show that saccharin or sucrose fading is not necessary when using alcohol-preferring animals.^11,12

Although the OSA paradigm has been used routinely in the context of alcohol consumption, recent research usually does not deepen in methodological aspects or the development of the operant behavior itself. The aim of this paper is (i) to describe and statistically analyze the basic steps of operant alcohol self-administration under a saccharin-fading procedure, (ii) to study the variables that affect the consumption through this paradigm, and (iii) to analyze the analogy with another paradigm commonly used as a model of voluntary alcohol consumption, to facilitate the understanding and performance of this paradigm.

METHODS Animals

Forty adult male Wistar rats (Harlan, Barcelona, Spain) were used to carry out the behavioral testing; we used male Wistar rats due to the abundance of information about the voluntary consumption of alcohol and other behavioral variables in these animals. Rats were weighted 225-250 g at the start of the experimentation and were housed in groups of four per cage (except as specified in the succeeding texts) in a specific pathogen-free and temperature-controlled and humidity-controlled environment (21 ± 1°C), on a 12-hours reverse light/dark cycle (lights off at 09:00 AM). Experimental sessions were performed during the dark phase. Food and water were available ad libitum except as specified in the succeeding texts. The animals were habituated to our facilities for 10 days before the beginning of the experimentation. All research was conducted in strict adherence to the European Directive 2010/63/EU and Royal Decree 53/2013 on the protection of animals used for scientific purposes. Animal studies are reported in compliance with the ARRIVE guidelines.¹³ The Ethics Committee of the Faculty of Psychology of the Complutense University of Madrid approved the study. All efforts were made to minimize animal suffering and to reduce the number of animals used.

Operant apparatus

The operant alcohol sessions were conducted in 12 operant chambers (Med Associates Inc.). The chambers were equipped with two retractable levers located 7 cm above a grid floor on either side of a drinking reservoir, positioned in the center of the front panel of the chamber and 4 cm above the grid floor. The levers had a refractory period of 3 seconds and were counterbalanced to respond as the active lever (delivering 0.1 ml of solution) or as the inactive lever (with no consequences for the animal). Auditory or visual cues were not present at any time. Operant self-administration sessions were performed daily during 30 minutes under a FR1 schedule of reinforcement throughout all the experiments.

Behavioral procedure Operant training

Animals were water deprived for 16 hours per day during the first 3 days of operant training to improve motivation and maximize the learning. For the rest of the experiments, the animals had access to food and water ad libitum. During the operant training, rats received 0.2% (w/v) saccharin solution (Sigma-Aldrich). An advantage of saccharin over sucrose is that it lacks calories.¹⁴ During this period, the lever press on 10 or more times was considered as the acquisition of operant responding.¹⁵ When all animals acquired the operant response, saccharin solution was reduced to 0.1% (w/v), in order to initiate the alcohol introduction.

Alcohol introduction and concentration increase

To proceed to alcohol introduction, we followed a protocol described in previous papers by our group^16–19 with slight modifications. Once the animals showed a stable consumption of 0.1% saccharin solution, we introduced alcohol as follows (Figure 1): 0.1% saccharin and 2% alcohol for three sessions (days), 0.08% saccharin and 4% alcohol for three sessions, 0.06% saccharin and 6% alcohol for four sessions, 0.04% saccharin and 8% alcohol for four sessions, 0.02% saccharin and 10% alcohol for four sessions, and finally, 0% saccharin and 10% alcohol for the rest of the sessions. Alcohol solutions were prepared from 96% alcohol (Alcoholes Aroca S.L.). During this period, we measured both the number of lever presses (ie, responses) and the consumption measured in gKg⁻¹ body weight (animals were weighted daily). Likewise, the progress of lever presses within operant session and the responses/rewards ratio were analyzed.

FIGURE 1. Behavioral procedure scheme. During training period, rats were trained to press the lever in the operant chambers to obtain 0.1 ml of a 0.2% saccharin solution. Once the animals acquired the operant response, the saccharin percentage was decreased to 0.1% to begin the introduction of the alcohol concentrations (which was accompanied by the reduction of the saccharin concentration). When the animals finished the alcohol concentration increase, a stabilization of consumption was carried out for 14 days and then there was an alcohol deprivation period for 7 days. In relapse period, animals were reintroduced into the operant chambers. Finally, the “Drinking in the dark-multiple scheduled access” paradigm was applied for 5 days in order to study the analogy between Operant Self-Administration with another paradigm commonly used as a model of voluntary alcohol consumption. DID-MSA, drinking iIn the dark-multiple scheduled access; EPM, elevated plus maze.

Alcohol deprivation

After the alcohol concentration increase and, once the animals showed a stable consumption of 10% alcohol solution (14 days), there was a period of alcohol deprivation for 7 days, in which the animals were not introduced in the operant chambers.²⁰ After that period, the animals were reintroduced back to the operant chambers, and the relapse-like behavior was evaluated over a period of 7 days. The consumption during these days was compared with the average consumption of the last 4 days before the deprivation (baseline). The reason for performing the alcohol deprivation period was that it is a procedure commonly used to assess the potential effect of several substances on relapse, so we wanted to assess whether consumption after such deprivation could be influenced by other variables, such as basal anxiety of animals. The percentage change in consumption during this period was analyzed with respect to the baseline level (Percentage change in Consumption = Consumption during Relapse × 100/Consumption during Baseline). Likewise, the relationship between percentage change in consumption during this period and the basal level of anxiety-related behavior shown by the animals was analyzed (see below).

Drinking in the dark-multiple scheduled access

To study the analogy between OSA with another paradigm commonly used as a model of voluntary alcohol consumption, we used the “Drinking in the dark-multiple scheduled access” (DID-MSA) paradigm.^21,22 Briefly, animals were presented two identical bottles in their home cages, one containing alcohol (10%) and another containing water. Daily, the procedure involved three 1-hours access sessions during the dark cycle, with the first session initiated at lights out and each subsequent period of access separated by 2 h of alcohol deprivation (ie, 9-10:00 h, 12-13:00 h, and 15-16:00 h). The procedure was carried out for 5 days. The bottles containing the alcohol and water were weighed before and after each 1-hour consumption period. The alcohol was renewed every day, to avoid its evaporation. The loss of liquid by dripping was controlled with two control bottles. The position of the bottles was randomly switched every day. The animals were weighed daily.

To assess consumption through this paradigm, animals were individually housed. The assessment of alcohol consumption through DID-MSA started 1 week after individual housing, to facilitate the habituation of the animals. During this week, animals were introduced into the operant chambers to avoid a posterior alcohol deprivation effect in the DID-MSA. It is important to note that there were no differences in consumption during this week (individual housing) and the previous days. Food and water were available ad libitum. To analyze the relation between consumption in OSA and DID-MSA, we used the average value of the last 7 days in OSA and the average of the 5 days of DID-MSA.

Elevated plus maze

During the operant training, 6 hours after self-administration session, basal anxiety-related behavior was evaluated using the “Elevated Plus Maze” paradigm,²³ one of the main methods for assessing anxiety-related behaviors responses in rodents. The apparatus was placed on 65-cm-tall legs and consisted of four arms (50 × 10 cm) forming a cross, two of them were closed and had 30 cm walls. The experiments were conducted in a quiet, dimly lit (15 lx) room. At the beginning of each measurement, the rat was placed in the center of the apparatus (10 × 10 cm) facing an open arm and the animal could explore the apparatus for 5 minutes. The experiment was recorded with a camera above the maze, and the following parameters were analyzed: time spent in the open and closed arms, and the number of open and closed arms entries. An arm entry was counted when the rat had all four paws on that arm. The percent of time spent in the closed arms (time spent in closed arms/total time spent in arms × 100) served as the measure of anxiety-related behavior.²⁴ The correlation between this behavioral variable and the consumption of the different solutions tested in the study was analyzed. In addition, the relationship between baseline anxiety and the percentage change in alcohol consumption during ADE was analyzed to study whether this variable could modulate the analysis performed on the consumption of the different solutions tested.

Statistical analysis

Data from Figures 2B, 3A–C and 5A were analyzed by repeated measures ANOVA, and the results were followed by Bonferroni post hoc test. Data from Figure 3D were analyzed by two-way mixed ANOVA (within subjects: interval; between-groups: drinking solution), and the results were followed by Tukey’s post hoc test in each interval. Correlations from Figures 4, 5B and 6 were determined by Pearson’s correlations analysis. A significance level of P < 0.05 was applied to all ANOVA statistical analyses. The SPSS statistical software package (version 20.0) for Windows (Chicago, IL, USA) was used for all statistical analysis.

FIGURE 2. Learning rate and stabilization of the operant presses throughout the training period. Rats received 0.2% (w/v) saccharin solution (n = 40). The pressure of the lever on 10 or more times was considered as the acquisition of operant responding. A, Data represent the percentage of animals, which acquire the operant response throughout this procedure by training day. On the sixth training day, 50% of the animals acquired the operant response. B, Data represent the stabilization of presses after the acquisition of the operant response. Day1 lever presses refer to the average of presses on the first day that each animal acquired the operant task. Rats increased their saccharin intake until day 8, and the intake remained stable over the next days. Mean ± SEM of cumulative intake for 30-min operant session are shown. **P [less than] 0.01, ***P [less than] 0.001 compared with Day1.

FIGURE 3. Consumption during alcohol concentration increase. Once the animals showed a stable consumption of 0.1% (w/v) saccharin solution, we introduced alcohol as described previously (n = 40). A, Data represent the progress of alcohol lever presses in each concentration of alcohol. The increase in alcohol concentration leads to a decrease in lever presses during the operant self-administration session. Mean ± SEM of cumulative consumption for 30-min operant session are shown. ***P [less than] 0.001 compared with 2% alcohol concentration. B, Data represent the progress of daily alcohol consumption measured in g·kg−1 body weight, in each concentration of alcohol. From the concentration of 4% alcohol, the consumption shows a greater stabilization. Mean ± SEM of cumulative consumption for 30-min operant session are shown. **P [less than] 0.01, ***P [less than] 0.001 compared with 2% alcohol concentration. C, Data represent the responses/rewards ratio in each solution. A significant increase in this ratio was observed as the percentage of alcohol increased. Mean ± SEM of responses/rewards ratio during 30-min operant session are shown. **P [less than] 0.01, ***P [less than] 0.001 compared with the last 4 days of saccharin consumption. D, Data represent the lever presses within operant session for the different solutions. At higher alcohol concentrations, animals tended to accumulate most of their presses in the initial period of the session. ***P [less than] 0.001 represents statistically significant differences between high concentrations of alcohol (6%-10%) and saccharin. gkg-1, grams per kilogram of body weight.

FIGURE 4. Consumption correlations between solutions. Figures represent the correlations between the consumption (lever presses) of saccharin solution used for training and each of the alcohol solutions introduced (n = 40). A positive correlation is observed with all concentrations except the last two (0.02% saccharin and 10% alcohol; 0% saccharin and 10% alcohol). Positive correlations are shown in blue. r: Pearson correlation coefficient.

FIGURE 5. Effect of 7-day alcohol deprivation on the consumption during relapse period (n = 40). A, Data represent consumption (g·kg−1 body weight) during 14 days of alcohol consumption stabilization. The highest value of consumption was 0.96 g·kg−1 and the lowest 0.73. B, Data represent consumption (g·kg−1 body weight) during 7 days of relapse period, compared with the average consumption of the last 4 days before the deprivation (baseline) (n = 40). During the first 2 days of the relapse period, there is an increase in consumption (alcohol deprivation effect). Mean ± SEM of cumulative consumption for 30-min operant session are shown. ***P [less than] 0.001 compared with baseline. C, Data represent Pearson’s correlation between baseline anxiety (evaluated during the operant training with saccharin) and the average of the percentage of change in consumption during the first 2 days of relapse period. gkg-1, grams per kilogram of body weight; 7-d, seven-day period; r, Pearson correlation coefficient.

FIGURE 6. Pearson’s correlation between the average value of the last 7 days of OSA and the average consumption of the 5 days of DID-MSA (g·kg−1 body weight) (n = 40). A moderate positive correlation is observed between both paradigms. DID-MSA, drinking iin the dark-multiple scheduled access; gkg—1, grams per kilogram of body weight; OSA, operant self-administration; r, pearson correlation coefficient.

RESULTS Training and acquisition of operant response

Figure 2 shows data from the operant training process. During this process, 40 rats received 0.2% (w/v) saccharin solution. Figure 2A shows the percentage of animals, which acquire the operant response throughout this procedure. On the first day of training, around 30% of the animals pressed the active lever more than 10 times during the 30-minutes operant session, which is the number of presses we considered as learning.¹⁵ This percentage increases to 50% on the sixth day and reaches 100% after 25 days of training.

Regarding the stabilization of consumption after the acquisition of the operant response (Figure 2B), Day1 lever presses refer to the average of presses made on the first day that each animal acquired the operant task. As soon as the animals acquired the operant response, there was an important increase in their consumption (repeated measures ANOVA: days, F [8, 311] = 11.590 P < 0.001) and statistically significant differences were observed from the third day compared with Day1. In addition, the Bonferroni’s post hoc analysis showed that there was a stabilization of consumption after 1 week. That is, the consumption observed on Day8 shows statistically significant differences with respect to each of the previous days (data not shown); however, during the interval Day8-20 no significant difference was observed.

Consumption during alcohol concentration increase

Once the animals showed a stable consumption of 0.1% saccharin solution, we introduced alcohol as described in methods section. Figure 3A shows the progress of lever presses in each concentration of alcohol. As expected, the increase in alcohol concentration caused a clear decrease in lever presses during the operant self-administration session (repeated measures ANOVA: concentration, F [2, 822] = 316.689 P < 0.001). There were statistically significant differences between all the concentrations, except between the last two (0.02% saccharin +10% EtOH and 0% Saccharin +10% EtOH, P = 0.131).

Figure 3B shows the progress of daily alcohol consumption measured in g·kg⁻¹ body weight, in each concentration of alcohol. As in the previous analysis, the change in alcohol concentrations led to an effect in alcohol consumption (repeated measures ANOVA: concentration, F [3, 698] = 23.472 P < 0.001). There was a significant increase in consumption at all concentrations compared with the first concentration. Also, there was a significant decrease in consumption during the last concentration, compared with previous concentrations (4, 6, and 8% alcohol). Figure 3C shows the responses/rewards ratio in each concentration of alcohol, compared with that observed during the last 4 days of saccharin consumption. We observed a significant increase of that ratio as the alcohol percentage increased (repeated measures ANOVA: concentration, F [3, 959] = 15.091 P < 0.001).

The progress of the lever presses within the operant sessions is depicted in Figure 3D. The ANOVA revealed a significant interaction between the solution administered and the session intervals (two-way mixed ANOVA: solution F (6, 273) = 61.883, P < 0.001; interval F (5, 1365) = 9938.460, P < 0.001; and interaction F (30, 1365) = 46.116, P < 0.001). During the last 4 days of saccharin consumption, animals showed a linear consumption pattern during the 30-minutes operant session. However, when different percentages of alcohol are introduced, animals tend to accumulate most of their lever presses in the initial period of the session. This accumulation of consumption at the beginning of the session increases with the percentage of alcohol, that is, the higher concentration of alcohol, the greater accumulation of presses in the initial intervals. In 8%-10% concentrations of alcohol, we observed that during the first 10 minutes of the session the animals have already reached 69%-73% of their total consumption. Also, in 6%-10% concentrations of alcohol, at 20 minutes of the session the animals exceeded 90% of their total consumption.

Correlation between consumption during training and the alcohol concentration increase

Figure 4 shows the Pearson’s correlations between the final saccharin solution (0.1%) and each of the alcohol solutions introduced. The last 4 days of saccharin consumption were significantly correlated with the initial alcohol concentrations (alcohol 2%: r = 0.77; P < 0.001; alcohol 4%: r = 0.60; P < 0.001; alcohol 6%: r = 0.66; P < 0.001; alcohol 8%: r = 0.44; P < 0.01). However, when the 10% alcohol percentage was introduced, this significant correlation disappeared (alcohol 10%-saccharin 0.02%: r = 0.295; P = 0.064; alcohol 10%-saccharin 0%: r = 0.240; P = 0.136). That is, the initial consumption of saccharin would not be a good predictor of the final consumption of high concentrations of alcohol.

Alcohol deprivation and consumption during relapse period

Once the period of alcohol concentration increase finished, a stabilization period was carried out for 2 weeks. Figure 5 shows the consumption during this period and subsequent consumption after 7 days of alcohol deprivation. First, Figure 5A shows that no significant variations were observed during the 14 days of stabilization (repeated measures ANOVA: days, F [5, 811] = 2.085 NS). Then, after 2 weeks of stable consumption, there was a period of alcohol deprivation for 7 days, in which the animals were not introduced in the operant chambers. Figure 5B shows the consumption (g·kg⁻¹ body weight) during relapse period, compared with the average consumption of the last 4 days before the deprivation (baseline). We observed that there is a significant effect during these 7 days (repeated measures ANOVA: days, F [4, 317] = 29.070 P < 0.001). A significant increase in the alcohol intake during the first 2 days of relapse period compared with the baseline and the rest of days was observed. This increase is known as alcohol deprivation effect (ADE). In addition, the average consumption during the 2 days in which the ADE was observed showed a significant correlation with the average consumption during the 2 weeks of stabilization (r = 0.720, P < 0.001). Figure 5C shows the Pearson’s correlation between basal anxiety-related behavior (evaluated during the operant training) and the average of the percentage change in consumption during the first 2 days of relapse period (g·kg⁻¹ body weight). Data showed a positive correlation between these variables (r = 0.364; P < 0.05), that is, the greater the basal anxiety-related behavior, the greater increase in consumption after a 7-day alcohol deprivation (during the evaluation of anxiety-related behavior, the average percentage of time spent in closed arms was 67%). On the contrary, basal anxiety-related behavior did not show statistically significant correlations with any of the solutions tested (sacc: P = 0.404; 2% alcohol: P = 0.505; 4% alcohol: P = 0.202; 6% alcohol: P = 0.857; alcohol 8%: P = 0.512; alcohol 10%_1: P = 0.470; and alcohol 10%_2: P = 0.931). Finally, basal anxiety-related behavior correlated negatively with the total number of entries in the open and closed arms (r = −0.411, P = 0.009), although this variable did not show a statistically significant association with any of the solutions tested.

Analogy between OSA and “Drinking in the dark-multiple scheduled access” (DID-MSA) paradigms

In order to study the analogy between OSA and another paradigm commonly used as a model of voluntary alcohol consumption, we used the DID-MSA paradigm. Figure 6 shows the Pearson’s correlation between the average value of the last 7 days of OSA and the average consumption of the 5 days of DID-MSA. Data showed that there is a positive correlation between these variables (r = 0.567; P < 0.01). The average consumption during the three accesses was 0.96 g·kg⁻¹ (1st access: 0.29 g·kg⁻¹; 2nd access: 0.30 g·kg⁻¹; 3rd access: 0.37 g·kg⁻¹; no statistically significant differences were observed between consumption in each of the accesses: 1-way ANOVA F [2, 119] = 3.025, NS).

DISCUSSION

The aim of this work was to analyze and review the process of acquisition of alcohol consumption through the paradigm of operant self-administration. We have described how the acquisition of the operant response occurs through saccharin fading and we have shown the association that this procedure has with the final consumption of alcohol. Likewise, we have analyzed the response of animals to an alcohol deprivation period and the association between this operant paradigm and another one frequently used (DID-MSA).

One of the main drawbacks of alcohol operant self-administration paradigm is the time required for the acquisition and maintenance of the operant response. Our data indicated that 25 days of training were needed for the acquisition of the operant response in all animals (40 rats). However, we observed that there were clear individual differences, since on the sixth day, more than 50% of the animals already pressed the lever more than 10 times in an operant session (Figure 2A). It is important to note that there are several studies in which the operant response is acquired without saccharin or sucrose fading.^6,25–27 However, saccharin fading could have high face validity, since it has been suggested that human drinkers usually drink sweet mixed drinks before progressing to stronger liquors.²⁸ Some strategies commonly used to maximize learning in this paradigm are to increase the time of the operant sessions,²⁶ restrict access to water 24 hours,²⁹ or to place the animals in the operant chambers for a 14 hours overnight sessions.⁹ One of the strategies that we often use in our laboratory when an animal does not learn the operant response is to place it in an operant chamber together with a cage partner who has acquired this response, in this way it is frequent that a vicarious learning takes place. Once the animals had acquired the operant response, they stabilized their consumption in a week (Figure 2B). This stabilization of consumption corresponds to around 55 ml kg⁻¹ body weight, which could indicate a satiety effect. Nevertheless, it is possible that this effect was due to the limited time of the sessions since animals show a linear consumption pattern during the 30-minutes operant session (Figure 3D). We think that saccharin/sucrose fading is especially useful in studies that do not use alcohol-preferring rats, since when introducing alcohol progressively, the probability that all animals consume alcohol is greater. In fact, in the present study, the lowest consumption during the final concentration of alcohol (10%) was 0.45 g·kg⁻¹.

As described in Figure 3A, during alcohol concentration increase, the animals reduced their lever presses in accordance with the increase in alcohol concentration. The last concentration (10% alcohol) led to a decrease of 71.3% compared with the average of presses obtained in the first concentration (2% alcohol). There were differences in presses between all the solutions, except between the last two, in which the consumption was stable. When we analyzed daily alcohol consumption in g·kg⁻¹ body weight (Figure 3B), consumption showed greater stability from the concentration of 4% alcohol (between 0.9 and 1.1 g·kg⁻¹). These data are similar to those obtained by other studies in which the percentage of alcohol was increased.^16,20,30 On the contrary, at the end of the alcohol concentration increase (10% alcohol) there was a significant decrease in consumption, similar to that described in other studies.³¹ Concerning this, the results depicted in Figure 3C could explain this data. This figure shows that the increase in the percentage of alcohol is accompanied by an increase in the responses/rewards ratio. In other words, in higher concentrations, more responses were given to obtain a single reinforcement. This may be due to a “compulsive” behavior of animals since they showed more presses during the refractory period of the levers (3 seconds). These data, together with the decrease in presses at high concentrations of alcohol, could explain the decrease in consumption measured as g·kg⁻¹ body weight in the last concentration.

Figure 3D shows the progress of presses within the operant session for the different solutions. During the last 4 days of saccharin consumption, animals showed a linear consumption pattern during the 30-minutes operant session. In contrast, when we increased the alcohol percentage of the solution, the animals tended to accumulate their consumption in the initial intervals of the operant session. It is an effect that has been previously reported.^20,32–35 This effect could be due to the fact that animals need less presses (and time) to reach the psychoactive effects of alcohol when its concentration increases. Our data indicated that when concentrations of 6%-10% of alcohol were used, animals reached 90% of their consumption during the first 20 minutes, so we could think that there was a “residual” consumption during the last 10 minutes. Regarding the relationship between previous saccharin consumption and final alcohol consumption (high concentrations) in OSA, we observed that there was no statistically significant correlation between these solutions (Figure 4). This indicates that saccharin consumption during the training period does not predict the subsequent consumption of alcohol.

The alcohol deprivation effect (ADE) was first described by Sinclair (1967).³⁶ Currently, it is a paradigm commonly used to study the effect of different molecules on voluntary alcohol consumption in rodents.^37–40 On the contrary, it has been described that anxiety can increase voluntary alcohol consumption, as well as the anxiolytic effect of alcohol.^41,42 In this line, data in Figure 5C show a positive correlation between the basal anxiety-related behavior of the animals and the percentage of change in consumption during the first 2 days of relapse period (ADE days). This data suggests that, in addition to alcohol-preferring animals, rat strains which are prone to express anxious-like behavior^43,44 could be very useful to study its relationship with alcohol consumption in this kind of paradigms, as well as the effects of potential treatments. The relationship between anxiety and alcohol consumption has been studied extensively. In fact, there is evidence about the efficacy of gabapentin, a calcium channel GABAergic modulator, on alcohol consumption. This treatment has also shown beneficial effects on anxiety through off-label clinical studies.⁴⁵ Finally, the consumption of alcohol during the last 7 days of OSA showed a moderate positive correlation with the subsequent consumption measured by the DID-MSA paradigm (Figure 6). This is an important result, since one of the strategies used to maximize learning in OSA paradigm is a prior access to alcohol under DID paradigm.²⁵ Both paradigms present clear differences: OSA requires prior learning of the operant task, so it has a certain cognitive component. Furthermore, this procedure is more associated with the reinforcing and motivational properties of alcohol in a specific context and within a restricted time of access. On the contrary, Drinking in the dark procedures are sometimes considered as a model of binge intake, since it is more related to an extended time of choice of drinking. In fact, under this type of protocol, it is common to expose the animals to alcohol for 12-24 hours or even present the alcohol solution as the only possibility for the animals to drink, thus forcing their consumption. Despite the indicated differences, the results obtained in this study show an important parallelism to evaluate alcohol consumption in animal models. The main limitation of this study is that part of the exposed data is known by experienced researchers. However, the articles usually do not deepen into these aspects and the objective of this work is to statistically analyze each of the phases that comprise this type of procedure. On the contrary, not all studies require a saccharin or sucrose fading for the acquisition of OSA, but this is the most common procedure when alcohol-preferring animals are not used. In this study, only a fixed ratio (FR) 1 is used. Although other ratio schedules are often used, FR was used due is the most common schedule of reinforcement. Finally, another limitation of the study is that blood alcohol levels are not analyzed, so all the data refer to exclusively behavioral variables.

Taken together, our results and analysis aimed to dissect the acquisition and maintenance of OSA behavior, as well as other related variables, to facilitate the understanding and performance of this paradigm. In this way, although OSA is often used to study voluntary alcohol consumption in preclinical models and leads to stable and moderate consumption levels, this paradigm can be combined with others to achieve high levels of consumption. As mentioned above, there are studies in which animals are trained to consume alcohol in a two-bottle-choice drinking procedure or are exposed to chronic alcohol vapor inhalation in order to induce dependence.⁴⁶ These types of procedures, as well as the use of alcohol-preference animals, make OSA a paradigm with multiple alternatives. For future complementary studies, it would be interesting to carry out studies using groups of animals exposed to other ratio schedules, as well as modifying the alcohol introduction and concentration increase process, to find the ideal protocol to reach stable levels of alcohol consumption.

AUTHOR CONTRIBUTIONS

JCC, EG, FRDF, and JALM were responsible for the study concept and design. JCC, VEA, KB, LSR, and PDG contributed to the acquisition of animal data. JCC, JAMG, EG, PO, and JALM assisted with data analysis and interpretation of findings. JCC drafted the manuscript. All authors provided critical revision of the manuscript for important intellectual content. All authors critically reviewed the content and approved the final version for publication.

CONFLICT OF INTEREST

The authors declare no competing financial interests.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available in the supplementary material of this article.

FUNDING INFORMARTION

This work was supported by National Plan on Drug Abuse, Ministerio de Sanidad of Spain (grant PNSD2018-050 to JALM), the Fondo de Investigación Sanitaria (Red de Trastornos Adictivos, FEDER, RD16/0017/0008 to JALM; RD12/0028/001 to FRDF) and the Instituto de Salud Carlos III (Sara Borrell research contract CD17/00125 to VEA).

ETHICAL APPROVAL

The Ethics Committee of the Faculty of Psychology of the Complutense University of Madrid approved the study. All research was conducted in strict adherence to the European Directive 2010/63 / EU and Royal Decree 53/2013 on the protection of animals used for scientific purposes.