Alcohol is one of the most used substance of abuse. In western countries around 80% of the adult population consumes alcohol1. Although the majority can consume alcohol in a controlled manner some individuals develop problematic levels of alcohol consumption. The individual differences in alcohol consumption form a substantial part of the problem. Detailed analyses are warranted to grasp the mechanisms that underlie the high degree of variability between individuals to develop AUD.
Although alcohol and other substances are quite different in their chemical structure and their primary neurobiological mechanisms, the characteristics of alcohol addiction are comparable to addictions to different substances of abuse. AUD and other SUDs are disorders consisting of a cluster of cognitive, behavioral, and physiological symptoms that for a large part indicate that the individual continues using the substance despite significant substance-related problems2. The Diagnostic and Statistical manual of Mental disorder (DSM V) encompasses eleven criteria for SUDs and is applicable for all the different substances. These criteria can be divided in four categories: impaired control over substance use, social impairment due to substance use, risky use (physical consequences of the substance use), and pharmacological criteria (developed tolerance and withdrawal symptoms after a period of abstinence)(table1). People suffering from a mild form of AUD or other SUD are diagnosed with at least two of the eleven DSM criteria, a person suffering from a moderate form of AUD or other SUD is diagnosed with four to five of the criteria and a severe form of AUD or other SUD is diagnosed with at least six or more criteria2.
Using these criteria a relatively large amount of individuals is diagnosed with an alcohol use disorder or AUD (4.1% of the adult population worldwide)1. From all the different addictive substances, alcohol is considered the most harmful substance for the user and its environment3. In 2012 alcohol accounted for 5,9% of all global deaths1. Despite the fact that alcohol use forms a serious health burden on an international scale, there is still a lot unknown about the neurobiological effects of alcohol and the mechanisms that determine the development of AUD. Although many studies addressed questions regarding the mechanisms that drive the development of AUD, focusing largely on reward related, craving related and abstinence related characteristics of addiction4, clinical treatment methods for AUD are still limited in their effectivity and number5. It is therefore essential to understand all and perhaps other aspects, which could create new insights and possible targets for improved treatment of AUD. Good animal models can be essential in accurately examining these different aspects of AUD. This gain of knowledge could eventually lead to more efficient treatments for people suffering from severe alcohol use disorders.
Over the past 10 years, different research groups have developed preclinical models to implement the discussed DSM criteria in animal studies. To that aim, specific DSM criteria for AUD and other SUDs have been translated into behavioral equivalents in rodents (table 2)6.
Studying behavioural equivalents of typical characteristics for people suffering from an AUD or other SUDs in animals will help to increase our knowledge of the mechanisms that underlie specific behavioural aspects of alcohol addiction in human beings6.
Alcohol use disorders & loss of control
A central hallmark of alcohol use disorders in human beings is loss of control over alcohol use, which is probably most reflected in the persistent seeking and drinking of alcohol despite adverse legal, health, economic, and societal consequences7. This central hallmark of AUD indicates the importance of loss of control and compulsivity, characterized by continued seeking or drinking despite adverse consequences, in AUDs8,9. Perhaps current treatment methods lack effectivity due to the fact that clinical research is not directed at restoring control over alcohol use, but rather focuses on reducing reward and craving. Preclinical studies still have not managed to demonstrate individual differences in voluntary alcohol consumption in non-alcohol preferring rats and accurately link these differences with differences in loss of control.
Therefore, it is essential to focus more on loss of control and compulsivity and individual differences in alcohol consumption. This could increase our knowledge of this disorder considering behavioural aspects of alcohol addiction. The related DSM criteria and the described central hallmarks of alcohol use disorders indicate that loss of control and compulsivity are important processes underlying alcohol addiction, as I have described previously for substance use in general (DSM-V criteria 1,2,3,4,6,8 and 9 in table 2). Considering the major contribution of loss of control to SUD and AUD and given the fact that current treatment strategies are aimed at reducing reward or craving, but do not focus on restoring control, it is essential to capture this aspect in preclinical models7.
Preclinical ‘loss of control’ models
Animal models that have been designed to incorporate loss of control and compulsivity in substance use a central hallmark indicating loss of control over substance use; ‘continued seeking or taking of the drug despite the knowledge of the adverse consequences’8. Animal models that capture compulsivity and loss of control typically use punishment paradigms to model substance use or substance seeking despite adverse consequences. The most widely-used adverse consequences within these preclinical settings are mild footshocks (0.2-0.4mA)10,11 and quinine addition12-14. Several studies showed that rats and mice develop, after escalation of substance use, aversion-resistant intake of the substance7,10,12,15. Aversion resistant intake of the substance indicates that animals continued to seek/consume the substance despite a bitter taste, footshocks or footshock-associated cues7.
There were several studies conducted on different substances mainly focusing on cocaine and alcohol8. I will focus on two important models considering loss of control and compulsivity in substance use: the quinine resistant model and the conditioned suppression model where either an aversive taste or a warning signal were used to assess the tendency of animals to continue their alcohol consumption patterns despite a negative consequence. The quinine resistant model in rodents was introduced by Wolffgramm in the 1990’s. As such, this was the first model considering loss of control in alcohol addict ion in rats13,16. He demonstrated that rats with an extensive history of alcohol use become insensitive to the bitter tastant quinine that was added to the alcohol solution, indicating aversion-resistant alcohol intake that may be interpreted as loss of control over alcohol use. Subsequently, several studies observed that both rats and mice developed a similar form of insensitivity to quinine after escalation of alcohol use in intermittent access paradigms12,14. Additively a recent study observed that in an intermittent access paradigm individual differences in the sensitivity for quinine can be observed17. This recent study showed that rats which consumed more alcohol in the alcohol drinking sessions (high drinkers) are less sensitive for quinine addition to the alcohol solution than the rats which consumed less alcohol (low drinkers). This indicates individual differences in aversion-resistant alcohol intake which may indicate differences in loss of control over alcohol use. Another model for loss of control is the conditioned suppression model in rats 10, where suppression of cocaine seeking in the presence of a warning signal is used to assess compulsive cocaine seeking10,11. In this model, rats are trained to press for cocaine in an operant task over a certain period of time. After the operant task the animals underwent fear conditioning sessions to achieve a tone-footshock association. This tone-footshock association was subsequently used to show that rats with an extended history of cocaine exposure, compared to rats with a limited history of cocaine self-administration, no longer suppressed their cocaine seeking behavior in the presence of a warning signal that is now represented by the tone. The conditioned suppression model has also been used to assess loss of control over alcohol use and compare limited and extended alcohol exposure18. In this model rats were exposed to the alcohol in an intermittent every other day model. After these drinking sessions animals with a comparable intermediate alcohol consumption level (medium drinkers) were selected for the conditioned suppression. After alcohol exposure the animals were exposed to the same experimental procedures as described in the cocaine study. In this recent study in our laboratory, it was shown that rats suppress their alcohol seeking behavior after limited exposure to alcohol in the presence of a warning signal represented by the tone. However after an extended period of alcohol exposure animals no longer suppressed their alcohol seeking behavior in the presence of a warning signal represented by the tone (Lesscher et al., unpublished findings). These data show that prolonged exposure to cocaine and alcohol leads to loss of control over cocaine and alcohol use, showing remarkable similarities to the quinine resistance described in the previous section.
Aim of this study – Alcohol and conditioned suppression
The primary aim of this study was to determine whether individual differences in alcohol intake predict differences in (loss of) control over alcohol seeking. To address this we used conditioned suppression of alcohol seeking as a measure for control over alcohol seeking. The designed approach is novel on several levels. For this study we used Lister Hooded rats that display substantial individual variation in alcohol intake, such that we can discern subgroups of low and high alcohol drinking Lister Hooded rats19. Furthermore, we applied for the first time conditioned suppression of alcohol seeking in these rats to determine individual differences in their control over alcohol seeking that we related to their initial level of alcohol consumption, comparing low and high alcohol drinking rats.
The results of this study will contribute to our knowledge of the factors that determine loss of control over alcohol use.
Neurobiological circuits underlying loss of control
It is intriguing to consider that perhaps modulation of neural circuits may help to restore control over alcohol seeking, first in laboratory animals and perhaps in the future also in humans using technologies such as deep brain stimulation20 or transcranial magnetic stimulation21. In order to achieve that, it is important to examine the neurobiology and neural circuits involved in control over alcohol seeking behaviour. Several brain regions have been implicated in the development of SUD, although the exact neural circuits in SUD and more specifically in loss of control over cocaine and alcohol seeking are not fully understood. Figure 1 visualizes the complexity of the wide spectrum of neurological pathways and brain regions involved in addiction22. Clearly there are multiple brain regions that contribute to different aspects of substance addiction. The ventral areas of the striatum especially the nucleus accumbens are extensively studied in substance and reward addiction8. More recently, the amygdala, the prelimbic cortex and the dorsal areas of the striatum have been proposed as important regions in the development of loss of control over drug use and the development of drug addiction8,23,24. Loss of control over substance use is already mentioned as an important characteristic of substance addiction. There are several theories considering the underlying neurobiological circuits and processes involved in the loss of control over substance use, two of which will be focused on in this study. One theory suggests the essential role of the dorsal striatal circuit and the other focusses on the possible role of the prelimbic cortex in loss of control over substance use (figure 1).
Figure 1: based on Everitt et al. 200625. (a) Key connectivities in human brain (b) Limbic cortical-ventral striatopallidal circuitry. The study related areas are encircled with red. (i) Processing of conditioned reinforcement and delays by basolateral amygdala and of contextual information by hippocampus. (ii) Goal-directed actions involve interaction of prefrontal cortex with other structures, possibly including nucleus accumbens but also dorsomedial striatum. (iii) ‘Habits’ depend on interactions between prefrontal cortex and dorsolateral striatum. (iv) ‘Executive control’ depends on prefrontal cortex and includes representation of contingencies, representation of outcomes and their value and subjective states (craving and, presumably, feelings) associated with drugs. (v)Substance craving involves activation of orbital and anterior cingulate cortex, and temporal lobe including amygdala, in functional imaging studies. (vi) Connections between dopaminergic neurons and striatum reflect ‘spirals”serial interactions organized in a ventral-to-dorsal cascade. (vii) Reinforcing effects of drugs may engage stimulant, pavlovian-instrumental transfer and conditioned reinforcement processes in the nucleus accumbens shell and core and then engage stimulus-response habits that depend on dorsal striatum. Green/blue arrows, glutamatergic projections; orange arrows, dopaminergic projections; pink arrows, GABAergic projections; Acb, nucleus accumbens; AMG, amygdala; BLA, basolateral amygdala; CeN, central nucleus of the amygdala; VTA, ventral tegmental area; SNc, substantia nigra pars compacta. GP, globus pallidus (D, dorsal; V, ventral); Hipp, hippocampus; mPFC, medial prefrontal cortex; AC, anterior cingulate cortex; OFC, orbitofrontal cortex; VS, ventral striatum; DS, dorsal striatum; Thal, thalamus.
The dorsal striatal circuitry and addiction
Several studies have implicated an essential role of the dorsal striatal circuit in loss of control and compulsivity for alcohol24 and cocaine addiction26. Upon extended periods of substance self-administration, escalation of self-administration occurs and drug seeking becomes progressively more controlled by substance associated stimuli24. With extended exposure to substances of abuse but also to natural rewards such as sucrose, operant responding for that reward is thought to shift from goal-directed behavior towards a more habitual controlled behavior25,27. This means that the behavior is initially driven by the outcome (the reward) but with extended exposure becomes independent of the outcome, as is illustrated by insensitivity to devaluation. For example, when animals with limited alcohol exposure are satiated with alcohol prior to an operant session, they will show reduced alcohol seeking as the alcohol is now devaluated. However, rats with prolonged alcohol exposure continue to seek alcohol even though the alcohol is devaluated, indicating habitual responding. The dorsal striatal circuitry plays an important role in habit learning or stimulus response learning and goal directed or response outcome learning24,28-30. Within this circuitry the Dorsal Lateral Striatum (DLS) is more associated with the habit learning and the Dorsal Medial striatum (DMS) is more associated with goal directed learning. An operational definition of a habit is that the behavior continues even when the value of the resource is decreased (devaluation). There are studies which indicate that the changes in alcohol and cocaine seeking behavior can be directly linked with neurobiological changes in this brain region24,26. For example, in an alcohol study the seeking behavior for alcohol was reduced upon devaluation, through preceding alcohol consumption in the homecage, in animals with limited alcohol exposure while devaluation did not affect alcohol seeking in animals with extended alcohol exposure, reflecting habitual alcohol seeking at this stage. Inactivation of the dorsolateral striatum reduces this habitual alcohol seeking, i.e. rendering rats sensitive to devaluation again. Thus, DLS inactivation reduces habitual controlled behavior and shifts behaviour to becoming goal-directed again. This suggests that the behavioral shift from goal directed behavior towards habitual controlled behavior in individuals suffering from or at least developing substance use disorder could neurobiologically involves a shift from DMS towards DLS control over behavior24,26. A second aim of this study was therefore to start assessing the role of the DLS in control over alcohol seeking behaviour. We hypothesize that dorsal striatum inactivation in animals that display loss of control over alcohol seeking can restore control over their alcohol seeking behavior.
The prelimbic cortex and addiction
Another brain region that has recently been implicated in control over substance use is the prefrontal cortex. The prefrontal cortex is an important region for cognitive control. In rats the prelimbic cortex is a homologous region of the prefrontal cortex regulating cognitive control, including decision making and inhibitory response control. Recently two studies examined this specific brain region in control over cocaine seeking in rats using optogenetic brain manipulation23 and a pharmacological inactivation methods31. One study demonstrated that pharmacological inactivation of the prelimbic cortex reduced conditioned suppression over sucrose and cocaine seeking behavior indicating loss of control31. The optogenetic study demonstrated that neuronal stimulation of the prelimbic cortex, using channelrhodopsin, reduced cocaine seeking behavior indicating regain of control over cocaine use. Whereas neuronal inhibition of the prelimbic cortex, using halorhodopsin, increased cocaine seeking behavior indicating loss of control over cocaine use23. The findings of these studies indicate that manipulation of the prelimbic cortex in rats affects drug seeking behavior23,31. Therefore, a third aim of this project was to assess the involvement of the prelimbic cortex in control over alcohol seeking . We hypothesize that prelimbic cortex inactivation in animals that remain in control over alcohol seeking can induce loss of control over their alcohol seeking behavior.
In this project, we started to assess the involvement of the DLS and PrL in control over alcohol seeking using an optogenetic approach .
The optogenetic approach is a recently developed approach that uses microbial opsins, which can be activated by illumination, to specifically manipulate specific cells in specific brain regions with a high precision32. Microbial opsins are proteins which use light sensitization to alter neuronal ion flux resulting in changes in activity of the targeted neurons. The opsin-expressing virus is surgically infused into the targeted area. This study uses halorhodopsin for its hyperpolarizing characteristics. Halorhodopsin expression initiated by laser illumination used in this study cause neuronal hyperpolarization in the targeted areas33. After the homecage alcohol consumption, high and low alcohol drinking rats will be discerned and prepared for surgery. Halorhodopsin will be expressed in the PrL of low alcohol drinking rats while the high alcohol drinking rats will receive infusions of halorhodopsin in the DLS. We hypothesize that a reduction of neuronal activity in the dorsolateral area in the CS+High group can restore control over alcohol use represented by conditioned suppression of alcohol seeking upon presentation of the shock-associated tone (arrow A, figure 2). For the low alcohol drinking rats, we hypothesize a reduction of the neuronal activity in the prelimbic cortex will induce loss of control over alcohol seeking, as would be evident from a reduction in conditioned suppression of alcohol seeking behavior ( arrow B, figure 2).
Figure 2. This figure illustrates the activity of the PrL and the DLS in the low and high drinking subgroups and the corresponding effects on control over substance use and the effects on conditioned suppression of drug seeking behavior. The arrows indicate the suggested effects of the optogenetic brain manipulation affecting control over substance use and the conditioned suppression of drug seeking behavior (altered figure34).
To summarize, this study used the conditioned suppression model to examine individual differences in loss of control or compulsive alcohol seeking, an important hallmark of alcohol addiction. Moreover, this addiction model is combined with optogenetics to assess the involvement of the DLS and PrL in control over alcohol seeking. The latter studies are still in progress and therefore only preliminary results for this part of the project will be presented.
Material & Methods
Male Lister Hooded rats (Charles River) were used for this study (N = 47 for individual differences (24+23) and N = 48 for the optogenetics). Initially the animals were used to examine the effects of adolescent alcohol exposure on adulthood alcohol consumption described in the supplementary section(N=48). These effects were examined by comparing adulthood alcohol consumption between two subgroups; animals exposed to water during adolescence(N=24) and animals exposed to alcohol during adolescence(N=24). Eventually only animals exposed to water during adolescence were used in the conditioned suppression test and the freezing test (N=24). For the assessment of individual differences in loss of control in the conditioned suppression experiment we combined the intermittent every other day alcohol consumption data and the conditioned suppression test data from an earlier study with the same setup (Lister Hooded rats, n=23 ) with the data retrieved from the animals exposed to water in adolescence(N=24). Animals of the former conditioned suppression experiment were not naï¿½ï¿½ve and were tested in an alcohol experiment, a play deprivation experiment and a pharmacological experiment before these animals started with the conditioned suppression experiment. After the pharmacological experiment the animals were re-exposed to the alcohol to achieve baseline alcohol consumption again. After achieving baseline alcohol consumption high and low drinking subgroups were discerned and the animals completed the stages of the conditioned suppression model. In the earlier study no freezing test was performed.
Figure 3. Timeline of the assessment of individual differences using the conditioned suppression model and the effects of adolescent alcohol exposure(AdolWat vs. AdolAlc) on adulthood drinking behavior(supplementary section). The individual differences experiment(only using the subgroup exposed to water during adolescence(AdolWat)) was combined with an earlier conditioned suppression experiment(+ earlier study). However the freezing test was not performed in the earlier study. Second timeline indicates the progress considering the optogenetics study. The stages illustrated in red were not completed yet.
The following timeline shows the different phases of the experiment(figure 3). I will mainly focus on the conditioned suppression related results and methods. Results of the effects of adolescent drinking on drinking in adulthood are presented in a supplementary section.
The Lister Hooded rats for examining individual differences in the conditioned suppression model were derived from a larger study, assessing the relation between social play, adolescent alcohol intake and adulthood alcohol consumption. The rats for this experiment were therefore not naï¿½ï¿½ve due to exposure to other experimental setting. They weighed between 300 and 370g and were individually housed in Macrolon cages. They were housed in a climate controlled room (temperature 20’21’C, 55 ï¿½ï¿½ 15% relative humidity) under a reversed 12H dark light cycle (lights OFF 7:00; lights ON 19:00).
The 48 Lister Hooded rats for the optogenetics arrived with a weight between the 200 and 238 gram.
All rats received ad libitum food and water The animals were checked for health and welfare related problems by handling and weighing them every week. Experiments were approved by the Animal Ethics Committee of Utrecht University, and were conducted in agreement with Dutch legislation (Wet op de dierproeven, 1996) and European regulations (Guideline 86/609/EEC).
Intermittent-every-other day alcohol consumption paradigm
For alcohol consumption rats were allowed to drink alcohol (20% v/v) according to an intermittent-every-other day schedule12,14,35,36. The alcohol solution was prepared by diluting ethanol 99.5% ethanol (Klinipath) in tap water. Alongside the alcohol bottle, the rats always had access to water also. This intermittent-every-other day schedule has been shown to induce high levels of alcohol consumption and escalation of alcohol intake in rats19,35. During the first four weeks, rats had access to alcohol for 7 hours per day on Mondays, Wednesdays and Fridays during the dark cycle (ï¿½ï¿½9:00-16:00). Thereafter, the time of alcohol exposure was extended to 24 hours (ï¿½ï¿½9:00-9:00), again on Mondays, Wednesdays and Fridays for four consecutive weeks. The position of the experimental alcohol and water bottles on the cage was switched every session. Additively we switched the position of the water bottle on non-drinking days after each alcohol intake session. These switches were made to prevent preference for bottle position on the home cage. Before and after each alcohol intake session the weight (grams) of the bottles (water and alcohol) was established to determine water and alcohol consumption. Every week we replaced the bottle, the caps, the water and the alcohol solution with clean material and solutions.
After consumption high and low drinking subgroups were identified based on their alcohol consumption. Animals were ranked based on their alcohol consumption, the total sum of rank of all the sessions was used to discern the subgroups. A medium split was used to efficiently use all the animals in the study. Otherwise the medium drinkers (1/3 of the animals) were excluded from final analysis which reduced our number of animals drastically. In the optogenetics experiment the decision was made to use a tertile split. This way the medium drinkers could be used to optimize the experimental conditions for the low and high drinking subgroups.
After two months of alcohol consumption, the rats were trained in an alcohol self-administration task in operant chambers. The rats were trained to lever press for alcohol under a heterogeneous seeking-taking chain based schedule of reinforcement with eventually a random interval (RI) of 120s on the active lever.
The operant training starts with a Fixed ratio 1 schedule of reinforcement (FR1) to achieve the acquisition of a response towards the active lever. When the animal presses the active lever, both the active and inactive lever will retract. Together with the lever retraction the cue light illuminates and the liquid dipper will move upwards facilitating an alcoholic reward (0.1ml, 20% EtOH). The liquid dipper stays up for 10 seconds after the animal made a nose poke (head entry) in the receptacle. After this a new trial will start, i.e. both levers will be presented again, the cue light is turned off and the liquid dipper is lowered into the alcohol containing reservoir. The inactive lever presses are counted however do not result in a reward presentation. Every session lasted 30 minutes. Active lever presses, inactive lever presses and the number of rewards were measured. Additionally, actual alcohol consumption was recorded by weighing the alcohol containing reservoir before and after each session. The alcohol solution was refreshed between sessions to avoid confounds due to evaporation.
To ensure all animals acquired the task sufficiently, the rats had to earn at least 10 reward under the Fixed ratio 1 schedule of alcohol self-administration for three consecutive sessions to proceed to the random interval training. They started with a random interval of 5 sec (RI5), slowly increasing over 15 sessions through RI5 (3x), RI15 (3x), RI30 (2x), RI60 (2x) to RI 120 (at least 5x). Under this schedule of reinforcement, an active lever press will initiate a RI. During the RI, the animals can press the levers but responding on the active lever will not result in access to alcohol. However, when the animal makes an active lever press after the random interval is completed, this directly results in a reward. The lever pressing during the random intervals not resulting in a reward and can be interpreted as alcohol seeking behavior. The RI5 – RI60 sessions last 30 minutes while RI120 session lasted 60 minutes. Active lever presses during the random interval, inactive lever presses during the random interval, nose pokes during the random interval and the amount of earned rewards were measured. Similar to FR1 sessions, the alcohol containers were weighed before and after each session.
The earned rewards in the first 15 min of each random interval sessions were used to determine response stability. Random interval responses were considered stable if the amount of responses remained within a specified range (25% deviation) for three consecutive sessions. Response stability was checked to assure stable and accurate responding in the conditioned suppression test.
Tone (CS) footshock pairing
Once the animals reached stable responding under the RI120 schedule of reinforcement, the animals underwent tone-footshock pairing sessions. Tone-footshock pairing sessions were established in an altered 5-choice box with a plastic screen covering the receptacles. To prevent context-shock association the animals underwent three habituation sessions of 10 minutes in the altered 5-choice boxes. The tone shock pairing sessions lasted 40 minutes. The first five minutes are a lead-in period with no shock or tone exposure followed by 10 minutes of tone (85 dB, 2900 Hz) exposure (CS, figure 4) combined with random footshock exposure (US, figure 4). During these 10 minutes, the animals received 10 unpredictable footshocks (0.40 mA) with a duration of 1sec each. Subsequent to this tone/footshock phase, a 10 min phase was incorporated during which there is no tone present and there is no footshock exposure. After this another 10 minutes phase with tone + footshock associations as described is incorporated in the session. The sessions ends with a five minute lead-out period with no shock or tone exposure. These sessions are fear conditioning sessions to establish an association between a tone (CS) and unpredictable footshocks. Within the identified high and low drinking subgroups, rats were assigned to either the footshock group (CS+) or the control group (CS-) based on alcohol intake and active lever presses during the first 15 minutes of the final 3 RI120 sessions, thereby ensuring that CS- and CS+ groups consumed and responded equally for alcohol. Only the footshock group (CS+) was exposed to random footshock exposure, the control group was exposed to the tone without combined footshock exposure. This thus yielded four experimental groups: the low drinking tone footshock paired group (CS+Low group), the low drinking control group (CS-Low group), the high drinking tone shock paired group (CS+ High group) and the high drinking control group (CS-High group).
The control group underwent similar sessions in which they received similar tone exposure time, however they were not exposed to the footshocks (CS-).
Conditioned suppression test
After conditioning and preceding the actual conditioned suppression test the animals received two additional random interval 120sec sessions to ensure stable responding under the RI120 schedule of reinforcement. The conditioned suppression test was performed in the same operant chambers as the alcohol self-administration. During the conditioned suppression test the two levers are presented, however active lever presses did not result in an alcoholic reward. The 12 minutes conditioned suppression test is subdivided into 2 min time bins during which the tone is present (ON-period) or the tone is absent (OFF-period) etc. Active and inactive lever presses made during the ON and OFF periods were recorded for every animal and were analyzed to assess the extent of conditioned suppression of alcohol seeking in response to the footshock associated tone. The alcohol reservoirs were placed in the boxes to avoid reduced motivation of the rats to seek alcohol in the absence of the scent of alcohol.
After the suppression test only the animals of the most recent executed study (N=24) were tested in the freezing test to control for a similar level of tone footshock association between the subgroups of low and high alcohol drinking rats. Therefore, after the animals past the conditioned suppression test, the animals were re-conditioned to associate random footshocks with a tone. In this case, due to miscommunication, the plastic screens were not placed to cover the receptacles of the 5-choice boxes as they were for the initial tone-shock pairings. The following day the animals underwent a freezing test. The test was established in the same five choice boxes as the tone-footshock pairings. To avoid differences in context the plastic screen was not placed at this time either. The test lasted for four minutes, starting with a two minutes period with no tone (CS) present and ending with a two minutes period where the tone (CS) was presented. On top of the five-choice boxes an infrared camera was placed to record the behavior of the rats during the test. After the behavioral recordings freezing behavior (the absence of any movement other than breathing10) was scored (frequency and duration) from tape with Observer software (Noldus, Wageningen, NL).
For the optogenetics experiment, the animals of this separate batch were assigned to one of six different subgroups prior to surgery. The six different identified subgroups are lowCS- eNpHR3.0, low CS+ eYFPcontrol , lowCS+- eNpHR3.0, highCS- eNpHR3.0,highCS+ eYFPcontrol and highCS+ eNpHR3.0. Animals were ranked based on their alcohol consumption, the total sum of rank of all the sessions was used to discern the drinking subgroups(High,Low). Additively animals were assigned to either the footshock group (CS+) or the control group (CS-) based on alcohol intake and active lever presses during the first 15 minutes of the final 3 RI120 sessions(LowCS-, LowCS+, High CS-, High CS+). The control group and a subgroup of the footshock group were injected with the halorhodopsin expressing AAV virus(lowCS- eNpHR3.0, lowCS+- eNpHR3.0, highCS- eNpHR3.0,highCS+ eNpHR3.0). However the additional surgery entails the necessity to control for the aspecific effects of virus injections and optical stimulation on the animals. Therefore two extra control groups are identified to control for these potential aspecific effects, receiving a YFP expressing control virus in surgery(low CS+ eYFPcontrol , highCS+ eYFPcontrol).
The batch for the combined conditioned suppression and optogenetic approach also achieved alcohol escalation, learned the alcohol self-administration tasks (FR1 and RI5-120) and underwent tone/shock association sessions. The difference between the experiments is the brain manipulation during the suppression tests and the surgeries executed before the animals started with the alcohol self-administration task (figure 5).
Figure 5.The timeline of both experiments is shown in this figure. Differences between the both experiments are displayed and described in red (altered figure ref34).
Surgery (Virus injection and chronic fiber implantation)
For stereotaxic infusion of the appropriate opsin expressing AAV virus and implantation of optical fibers, the rats were anesthetized with ketamine (75mg/kg) and dexdomitor (0,25mg/kg). Subsequently, the rats were fixed into a stereotactic frame and the animals received virus microinjections (0.3 ï¿½ï¿½l eYFP control AAV virus or eNpHR3.0 halorhodopsin expressing AAV virus at a speed of 0.3 ï¿½ï¿½l/min) bilaterally through 30G microinjectors into the dorsolateral striatum (coordinates from Bregma: +2,8 AP, +1,6 ML, -3.1 DV) or the prelimbic cortex (coordinates from Bregma: +0,8 AP,+3,4 ML,-5.3 DV). The microinjectors were left in place for at least 1 minute after infusion to allow diffusion of the viral particles. The microinjectors were removed and subsequently, two chronic optic fibers were implanted bilaterally above the dorsal lateral striatum (coordinates from Bregma: +2,8 AP, +1,6 ML, -2.6 DV) or the prelimbic area (coordinates from Bregma: +0,8 AP,+3,4 ML,-4.8 DV). Chronic optic fibers were handmade in the lab described in box 137.
The chronic fibers were secured to the skull with antibiotic containing dental cement, supported by two small screws that were inserted into the skull. Immediately after surgery, the rats received a carprofen injection (5 mg/kg) for pain management and an Atipamezol injection (0.6 mg/kg) to reverse the anesthesia and induce recovery of the animals. The rats received an additional carprofen injection (5 mg/kg) 24 hours after surgery for pain management. Immediately after surgery the animals were placed in Individually Ventilated Cages (IVC) with HEPA filters to rule out the risk of potential AAV infection. The rats remained in the IVC housing for one week in a room with a normal day-night cycle (lights ON 7:00 and lights OFF 19:00). After one week the animals returned to the room where they were housed before the surgery, with a reversed 12H dark light cycle (lights OFF 7:00 and lights ON 19:00). Rats were weighed every day until their weight returned to pre-surgery levels. The animals were not tested for another week to allow full recovery from the surgery. After this week, the animals were trained to self-administer alcohol using procedures described above). After the animals achieved to learn the task proper at the random interval 120 seconds the same conditioned suppression test was executed as described earlier, however now in combination with optical stimulation.
To assess the role of the DLS and PrL in control over alcohol seeking, we combined optical inhibition with the previously described conditioned suppression test. For the optical stimulation the chronic fibers of the animal will be connected with a ceramic sleeve to an optical fiber directly attached to an optical patch cable (figure 6). This cable is connected via a rotary joint (Doric Lenses), through a counterbalanced lever arm to allow free movement of the animals in the operant chambers while connected, to the laser (MGL-III-532-100mW with TTL modulation BNC (1Hz-1kHz) and PSU-III-LED power supplies; Laser 2000). Every animal will be acclimatized to being connected and to light exposure before actual testing. During the earlier described tone ON periods the rats receive optical stimulation (8-12mW, 300sec, 532nm, PSU-III-LED) to achieve DLS and PrL inhibition by eNpHR3.0 (figure 7). Animals with the eYFP control virus receive optical stimulation however the optical stimulation does not affect neuronal activity due to an absence of halorhodopsin. Additively comparable pseudostimulation sessions will be executed. In the pseudostimulation sessions the laser is turned on but laser light is s blocked from entering the brain by inserting black foam-like material in the ceramic sleeve. Active and inactive lever responses are recorded during the conditioned suppression test as described previously. Following the suppression test, rats will tested for freezing behavior the same way described earlier.
Figure 7. The conditioned suppression test including optical (pseudo)stimulation during ON periods (tone present).
In the first experiment considering individual differences we used a total of 47(24+23) animals and in the second optogenetic experiment we used 48 animals. The data of the suppression test and the alcohol consumption data in the former conditioned suppression experiment (N = 23) were combined with the same data retrieved from the animals exposed to water in adolescence(supplementary) in this study(N = 24) for the first experiment(material & methods). For an accurate combination of the data considering the first experiment two sessions of the 7H sessions and three sessions of the 24 sessions were excluded from the analysis. In the second optogenetic experiment due to practical concerns we decided to discern three subgroups, a tertile split (low, medium, high). In the analysis of the second experiment two animals were excluded due to missing values for alcohol consumption (one high drinker and one low drinker). The alcohol intake and preference data in both experiments and active responses conditioned suppression data were analyzed by repeated measure ANOVAs with ON/OFF periods as the within-subjects factor and subgroup (Low or High) and conditioning (CS- or CS+) as between-subjects factors. Subgroups (High, Low) in the conditioned suppression test were analyzed separately by using an ANOVA. The freezing data were analyzed by an ANOVA. When appropriate, data were further analyzed by student t-tests. The data shown represent mean ï¿½ï¿½ SEM. Statistical analyses were performed using SPSS 22.0 software. Significance was accepted at P < 0.05. I used the gathered data from experiment one to illustrate the effect of the optical stimulation we expect to observe in experiment two. Results The data presented for experiment one are final data of this study combined with data from a previous study. The data from experiment two are data from an experiment which is still ongoing and are therefore not yet complete. Experiment 1 Individual differences in the conditioned suppression model Alcohol intake sessions The subgroups are identified based on individual differences in alcohol intake in the 24H sessions. A medium split was used for subgroup identification (low, high). The subgroups significantly differed in their average alcohol (20%EtOH) intake per session over all sessions [Fsubgroup*sessions(17,731) = 1.3518, p<0,01](figure 8). Additively the high drinking subgroup has a significant higher preference for the alcohol solution compared to the low drinking subgroup over all the sessions[Fsubgroup*sessions(17,731)= 2.881, p<0,01](figure 9). Figure 8. This figure illustrates the difference in average EtOH intake (g/kg) between the low drinking subgroup (n=22) and the high drinking subgroup (n=23) in the 7H sessions(9) and the 24H sessions(9) [Fsubgroup*sessions(17,731) = 1.3518, p<0,01].Significant difference within sessions between subgroups indicated by *(p<0,05) or **(p<0,01). Figure 9. This figure shows the preference in percentages for the alcohol solution compared to the water over the 7H sessions(9) and the 24H sessions(9). The high drinking subgroup(n=23) had a significant higher preference(%) for alcohol compared with the low drinking subgroup(n=22). Preference is calculated as the percentage of the amount of alcohol consumed relative to the total amount of fluid (water + alcohol) consumed [Fsubgroup*sessions(17,731)= 2.881 p<0,01] . Significant difference within sessions between subgroups indicated by *(p<0,05) or **(p<0,01). Alcohol self-administration In the last three of the random interval 120 of the animals the stability and the average active lever presses during the first 15 minutes were established. There were no significant differences observed between the high and low drinking subgroups. However there was a trend towards a higher average number of active lever presses for the higher drinking subgroup (t-test, p=0.0979)(figure 10). Figure 10. Average active lever presses during the first 15 min of the RI120. There are no subgroup differences between Low drinkers(n=23) High drinkers(n=24). There is only a slight trend(p=0.0979) towards a higher average active lever press count for the higher drinking subgroup. The suppression test The main and most important result of this study is shown in figure 11. This figure illustrates the active responses for the low drinking subgroup (A) and the high drinking subgroup (B) made during the first suppression test when the tone was present (ON) and when the tone was absent (OFF). Figure 11A illustrates the differences in active responses between the control group (CS-) and the fear conditioned group (CS+) within the low drinking subgroup. The figure displays an overall significant higher amount of active responses over the whole session in the low drinking control group (CS- low group) compared to the fear conditioned low drinking group (CS+ Low group) [Fcs*ONOFF (5,105)=9.732, p<0.01]. Between subjects there was also a significant CS subgroup effect(Fsubgroup(1,21)=9.732, p<0.01). No differences were found in responding in the ON or OFF periods in the low drinking subgroup[FONOFF(5,105)= 1.412,p>0,05].A significant higher amount of active responses was found on the first ON (p<0.01), the first OFF (p<0.01), the second ON (p<0.01) and the third ON (p<0.05) for the control group compared with the the fear conditioned group (CS+). Figure 11B illustrates the differences in active responses between the control group (CS-) and the fear conditioned group (CS+) within the high drinking subgroup. The variance in the data was not equal and therefore, the Greenhouse-Geyser corrected values are presented. The figure displays that there was no differences between the fear conditioned group (High CS+) and the control group (High CS) across ON/OFF phases of the conditioned suppression test [Fcs*ONOFF (5,211)=2.325, p>0.05]. There were also no differences observed between CS groups [Fcs(1,22)=2.897 p>0,05].
Figure 11. The amount of active responses (presses on the active lever) during periods when the tone was present (ON period) and when the tone was absent (OFF period). In figure 11A CS-Low (n=12) is significant higher in their active responses in the suppression test compared to the CS+ Low group (n=11) [Fcs (1,21)=9.732, p<0.01].In figure 11B no differences were observed comparing the active response of CS-High (n=12) and CS+High (n=12)[Fcs*sessions (5,211)=2.325, p>0.05]. Additively no differences between subjects were observed[Fcs(1,22)=2.897 p>0,05]. Significant difference within sessions between subgroups indicated by *(p<0,05) or **(p<0,01). The freezing test Analysis of freezing behaviour for low or high alcohol drinking rats revealed that there is a difference between the subgroups, dependent on the conditioning (CS+ or CS-) [Fsubgroup x CS(3,20)=12.643,p<0,01]. Both subgroups showed a significant increase in the % freezing of the observed time when the tone was presented (t-test, p<0.01). However, the CS+Low subgroup displayed a significant higher % freezing of the observed time when the tone was presented compared to the CS+ High subgroup [Fsubgroup(1,12)= 12.217,p<0,01]. These results indicate that among both subgroups (low, high) the fear conditioning of the tone combined with the footshock was effective, although the extent of freezing was greater in the Low drinkers(figure 12). Figure 12. This figure displays the percentage time freezing of the observed time in the freezing test. Both the fear conditioned (HighCS+ (n=6)and LowCS+ (n=6)) subgroups spend a significant higher percentage of the time freezing compared to the control subgroup (HighCS- (n=6) and LowCS- (n=6))[ttest, p<0.01].Additively this figure shows an interaction effect of subgroup and CS [Fsubgroup x CS(3,20)=12.643,p<0,01] Experiment 2 The combined conditioned suppression and optogenetic approach Alcohol intake sessions The optogenetics experiment, aimed to assess the involvement of the DLS and PrL in control over alcohol seeking, is not yet completed. The only available data for this experiment are the voluntary alcohol consumption data. Based on the alcohol consumption data, the animals were assigned to one of three subgroups (low, medium and high alcohol drinking rats) using a tertile split based on the sum of rank score for the 8 consecutive weeks of alcohol consumption. A tertile split was used due to earlier indicated practical reasons. The subgroups differed significantly in their level of alcohol intake over sessions (Fsessions*subgroup(46,966)= 9.038, p<0.01], figure 13) and in preference for the alcohol solution ([Fsubgroup*sessions(46,920)=4.135, p<0,01] , figure14). Figure 13. This figure illustrates the difference in average EtOH intake (g/kg) between the low drinking (n=14), medium drinking (n=17) and the high drinking subgroup (n=14) in the 24H sessions[Fsessions*subgroup(46,966)= 9.038, p<0.001]. Figure 14. This figure shows the preference in percentages for the alcohol solution compared to the water over the 12 24H sessions. The high drinking subgroup(n=13) had a signifcant higher preference(%) for alcohol compared with the low drinking subgroup(n=13) and the medium drinking subgroup (n=17). Preference is calculated by comparing the amount alcohol consumed to the amount of water consumed in percentages [Fsubgroup*sessions(46,920)=4.135, p<0,01 ] . The three groups were trained, after surgery, to self-administer alcohol in operant boxes under FR and RI schedules of reinforcement, ultimately working towards RI120 responding. During the course of operant training the animals were habituated to the attachment of an optical cable on their heads, first with a dummy cable in the homecage and ultimately with actual optical patch cords connected to lasers via rotary joints and ultimately also with pseudostimulations where the laser is turned on but laserlight was blocked from entering the brain. So far, we have assessed the effects of connecting the animals and pseudostimulation during alcohol self-administration under a 30 minutes session under a RI120 schedule of reinforcement. During these tests, the animals were observed and corrected when they stood up with the tendency to start grabbing the optical cable. The initial findings for medium drinkers who on three consecutive days received a pseudostimulation, then DLS inactivation and ultimately again a pseudostimulation show no effect of DLS inactivation on baseline responding under the RI120 schedule of reinforcement. The actual conditioned suppression test and inactivation of DLS or PrL in high and low alcohol drinking rats, respectively, has not yet been performed. Therefore I will illustrate the expected outcome of this study based on our hypothesis that reduced activity of the PrL and enhanced activity of the DLS contribute to loss of control over alcohol seeking. Expected results experiment 2 The following figure visualizes the expected effects after optical inactivation in low and high alcohol drinking rats. The expected results are integrated in the graphs from experiment 1 to visualize the hypothesized findings. In the fear conditioned low drinking subgroup (CS+ Low) we expect that inactivation of the neurons in the prelimbic area (PrL) will induce loss of control over substance use as apparent from a reduction in conditioned suppression over alcohol seeking behavior within this subgroup. The significant suppression of alcohol seeking in the conditioned low drinking subgroup (CS+ Low) compared to the control low drinking subgroup (CS-Low) as found in experiment 1 would disappear (Figure 15A). In contrast, we expect for the conditioned high drinking subgroup that inactivation of the neurons in the dorsal lateral striatum (DLS) would restore control over alcohol seeking as would be evident from conditioned suppression of alcohol seeking behavior within this subgroup. This would be observed from a significant reduction in active responses in the conditioned high alcohol drinking group (CS+ High) compared to the responding by the high drinking control group (CS- High) (figure 15B). Discussion The conditioned suppression model Despite the substantial amount of studies performed considering alcohol use disorders an effective treatment strategy and knowledge of the neurobiological processes underlying alcohol use disorders are still sparse5. Recently, aspects considering loss of control and compulsivity of alcohol and other substance use disorders are studied more extensively. Primarily because loss of control and compulsivity comprise several important DSM-V criteria. There is still a lot unknown considering these important criteria of the DSM-V for AUD. Therefore the development of a proper and valid preclinical model is essential for generating more knowledge of these aspects leading to new targets for clinical research and treatment. This preclinical model properly incorporates the related DSM criteria. Escalation of alcohol use is achieved in the alcohol sessions. The training for self-administration incorporates the motivational aspect in this model. Finally the suppression test gives us information considering resistance to punishment among the different animal subgroups. Studying these three behavioral equivalents 8/11 DSM criteria can be investigated in only one model (table 2). This study developed a comprehensive preclinical model for loss of control over alcohol use that is a key characteristic of alcohol use disorders (AUD). Our main finding is that individual differences (low vs. high) in alcohol intake are predictive for individual differences in conditioned suppression over alcohol seeking behavior. The fear conditioned low drinking subgroup displays a significant reduction in active responses to obtain alcohol in the presence of the aversive conditioned stimulus compared with the low drinking control group (figure 8). This demonstrates that the Low CS+ subgroup suppresses alcohol seeking behavior in the presence of a warning signal, indicating being in control over alcohol use. The fear conditioned high drinking subgroup does not differ
from the high drinking control group in active responses to obtain alcohol in the presence of the aversive conditioned stimulus. This shows that the High CS+ subgroup does not suppress alcohol seeking behavior in the presence of a warning signal, indicating developed compulsivity and loss of control over alcohol use. This proves that the conditioned suppression model is an excellent model to examine individual differences in alcohol intake and loss of control over alcohol use. The found results contribute to the earlier results found considering loss of control and substance use using the conditioned suppression model. In a former study we already established that the amount of time exposed to the alcohol is essential in the development of loss of control over alcohol use18. Additively a recent study of Spoelder et al. (unpublished) found individual differences between high and low drinking subgroups in aversion resistant intake indicating differences in loss of control17. Combined, these studies indicate that both individual differences in alcohol consumption and amount of time exposed to the alcohol are important predictors for loss of control over alcohol use. The used combined approach of the conditioned suppression model and the intermittent every other day alcohol drinking paradigm has several important advantages over the other models for the examination of individual differences in alcohol consumption and loss of control over alcohol use. One important aspect of the model is that we used a paradigm that induces substantial alcohol consumption in a voluntary binge-like pattern. Many studies have shown that rodents develop inflexible drinking behavior12,35 and clearly increase their drug intake over time12,35 by using the free choice drinking models. We used a variation on the free choice drinking model, a developed 2-bottle-choice drinking paradigm with intermittent access to alcohol19. The adapted intermittent-every-other day model and other free choice drinking models are a very valid, low cost and simple way to induce escalation of drug use in rodents. An important advantage of this paradigm over for example vapor38, liquid diet39 or continuous free choice bottle drinking models35 is the progressive escalation of alcohol intake which is very comparable with the development of alcohol addiction among human beings40. In this study there were clear subgroup differences observed in average alcohol intake and in preference for the alcohol solution. These individual differences in alcohol intake among the rats form an excellent representation of individual differences found among a group of human beings. Importantly, the individual differences in consumption also translate into differences in control over alcohol use, further increasing the validity and strengths of this approach. Another advantage of this model compared to the other models for loss of control and compulsive substance use is that the condition suppression model uses fear conditioning (tone-footshock association) to accurately model the adverse consequence in a preclinical setting. The use of a sufficient adverse consequence is essential in preclinical modelling the central hallmark of loss of control; ‘continued seeking or taking of the drug despite the knowledge of the adverse consequences’8. The quinine resistant models and other footshock based models use different ways to accurately model the adverse consequence. The quinine resistant model studies uses the addition of the bitter tastant quinine to model the adverse consequence. Using this model several studies demonstrated that time exposed to alcohol induces aversion resistant intake indicating loss of control12-14. Additively individual differences between high and low drinking subgroups in aversion resistant intake were found in a recent study of Spoelder et al. (unpublished) indicating differences in loss of control using the quinine resistant model17. Other footshock based models use unpredictable footshocks during the operant self-administration task to assess the effects of adverse consequences related to the seeking and taking of cocaine and/or alcohol15,23,41,42. The footshock exposure initially results in reduced responding in the operant task. However, after several sessions a subgroup of animals exhibit responding comparable to pre-shock conditions. This footshock-resistant intake model is the first model, which shows that a subgroup of animals displays footshock-resistant substance seeking. Two studies observed a small punishment resistant subgroup among the animals with an extended exposure to drug compared with no punishment resistant subgroup among the animals with a limited exposure to the drug15,42. Footshock-resistant cocaine/alcohol seeking could indicate loss of control. The main point of criticism considering the preclinical modelling of an adverse consequence in the quinine resistant models and other footshock based models is that these adverse consequences are directly combined with the consumption of the substance. The quinine resistant model observes resistance to aversion exclusively experienced during the consumption of the alcohol. And the other footshock based model exposes the animals to unpredictable footshocks only during the operant self-administration task15,23,41,42. Both models do not fully encompass the wide variety of adverse AUD related consequences among humans. Among humans there are diverse not direct consumption-related adverse consequences such as relationship problems, problems with family and friends problems at work or other problems where people suffering from an AUD have to cope with. People suffering from an AUD are regularly exposed to these stressful and problematic situations which are not directly consumption related. For a sufficient examination of the wide spectrum of loss of control related aspects it is better to use the conditioned suppression model. The conditioned suppression model uses fear as the adverse consequence without exposing the animal to the actual punishment (the footshock) or the pain related context in the operant task or during alcohol consumption. The tone-footshock association session is a single fear conditioning session to associate a footshock with a tone in a completely different context than the original setting. The fear conditioning session is performed to eventually create a adverse context or situation for the animals in the conditioned suppression test, not related to the consumption of alcohol. The use of fear to create an adverse context in the conditioned suppression model better reflects the wide and diverse AUD related aspects among humans. A second point of criticism concerns the setup of several studies using the footshock based model. These are several studies which examined animals which consumed comparable amounts of cocaine/alcohol. After this the animals were exposed to the unpredictable footshocks to identify footshock-sensitive and footshock-resistant animal subgroups. Following the alcohol exposure these studies used optogenetic brain manipulation among the footshock-resistant subgroup to observe differences in footshock resistant alcohol/cocaine intake23,41. However the selection of a specific group of animals biases the found results towards a specific punishment resistant subgroup. The footshock-resistant animals in these studies are selected based on resistance to punishment not based on vulnerability for the development of an AUD, making it less representative for the human situation. The conditioned suppression model uses fear conditioning to avoid interference of the effect of differences in sensitivity for the shock on the results. Different studies confirmed that the tone-footshock association affects all the animals. This is quite different compared to the specific selected subgroup which developed foothshock-resistant intake in the footshock based model10,11. This way all the subgroups identified based on alcohol consumption(Low,High) can be examined in the conditioned suppression for loss of control without interfering effects of individual differences in sensitivity for punishment. Efficacy of
tone-footshock association is controlled by measuring freezing behavior in a freezing test. This test is essential to check for differences in sensitivity for the footshock and check if the animals successfully associated the tone with the shock. In our study we checked for differences in sensitivity for the footshock among the animals by the integration of the freezing test in the model. In our study the freezing test confirmed that the animals in the targeted subgroups (CS+Low, CS+High) were all sensitive for the footshocks in the tone-footshock association observed in a clear difference within both subgroups in freezing behavior between the control group(CS-) and the fear conditioned group(CS+). The efficacy of the fear conditioning is represented by the significant difference found in the freezing behavior between the conditioned subgroups and the control subgroups when the tone is present. Additively found subgroup differences suggest that there are small differences in sensitivity for the punishment or reaction on the punishment which should be further investigated. However the within subgroup differences were the most valuable to control for the efficacy of this fear conditioning method in this model. This way the conditioned suppression model can examine control over alcohol use in all the animals without selecting for a certain group which is more punishment resistant. The quinine resistant model and the footshock model tell us more about the aversive-resistant drug intake directly related to the consumption of the compound. The condition suppression model adds another aspect by creating an adverse situation for the animal completely independent of the alcohol itself. Additively the conditioned suppression model uses fear conditioning in all the animals effectively and prevents interference caused by individual differences in sensitivity for the shock on the results. All the models should be considered as valid models in further research however one should take in account that the different models represent different aspects of alcohol use disorders. Additively I want to discuss an interesting minor difference found in our results compared to the conditioned suppression test results in the first study which developed this model10. The interesting aspect of the results we found is that the animals did not responded differently when the tone was on than when the tone was off during the conditioned suppression test. The CS+ Low subgroup suppressed alcohol seeking behavior over the whole session, also when the tone was not presented in the OFF periods. This is an essential difference compared to the results found by Vanderschuren in his cocaine study who only found conditioned suppression during ON periods in animals which were limited exposed to cocaine. A possible explanation of the different found results is that the animals exposed to the alcohol were so impressed and overwhelmed by fear following the tone that a period of two minutes was not sufficient enough for recovery from such a stressful and fearful event. And apparently the animals which were exposed to the cocaine were less impressed and overwhelmed by fear and recovered quite quickly from this fearful event. Possibly the differences between both experiments could be explained by individual differences between animals in experiencing such a fearful event. This could have resulted in a suppression of alcohol seeking behavior in both the ON and OFF periods compared to a suppression of cocaine seeking behavior in only the ON periods. More research is necessary to investigate this differences in conditioned suppression between the ON/OFF periods and for cocaine and alcohol. Despite this minor difference the validity of the conditioned suppression model to accurately measure individual differences in loss of control remains excellent as described above. A potential confound of this study is the usage of animals which were not naï¿½ï¿½ve. Before the study started the used animals were already tested for cognition and play behavior. Earlier exposure to different experimental conditions possibly could have influenced the behavior of the animals in this study. In short we demonstrate that the conditioned suppression model models individual differences in loss of control and compulsivity considering alcohol use disorders. The incorporation of several different DSM criteria combined with an excellent face validity on many levels show that this is an excellent preclinical model to examine individual differences in loss of control and compulsivity in AUDs. The combined conditioned suppression and optogenetic approach These results make it possible to proceed to a next step where an optogenetic approach will be integrated into the model to actually examine effects of brain stimulation in the areas of interest. Having shown that low and high alcohol drinking subgroups of rats differ in their level of control over alcohol use, we have also set important grounds for the analysis of involvement of the PrL and DLS in control over alcohol seeking. Within the next months the effects of PrL and DLS inactivation on control over alcohol seeking in low and high alcohol drinking rats will be established. The prelimbic area and the dorsal lateral striatum are the areas of interest considering loss of control and compulsivity. In a study designed by Chen they indicated that the preIimbic area could be an essential area in the neurobiological mechanisms underlying loss of control over substance use. They demonstrated that optogenetic inhibition of the prelimbic area induces loss of control over cocaine seeking in footshock-sensitive rats. In this study we want to observe whether the optogenetic inhibtion of the prelimbic area also induces loss of control over alcohol seeking in rats in the conditioned suppression model. Another study designed by Corbit indicated that the dorsolateral striatum(DLS) could be an important area considering the neurobiological mechanisms underlying loss of control over alcohol use. In this study they showed that pharmacological DLS inactivation induces a regained control over alcohol seeking behavior. In this study we want to observe whether optogenetic inhibition of the dorsolateral striatum induces a regain of control over alcohol seeking in the conditioned suppression model. These are the two main brain areas of interest in this study, both potentially essential areas in the neurobiological mechanisms underlying loss of control over alcohol use. Considering other found literature we expect to find an effect on alcohol seeking behavior of the performed brain stimulation. The hyperpolarizing effect of halorhodopsin considering the prelimbic area and the dorsal lateral striatum in the different subgroup could possibly affect loss of control over alcohol use (figure 2, figure 15). The neuronal inactivation of the DLS in the fear conditioned high drinking subgroup would result in a gain of control over substance use represented by a regain of conditioned suppression over drug seeking behavior. The neuronal inactivation of the prelimbic area in the low drinking subgroup would result in loss of control over substance use represented by an absent conditioned suppression of alcohol seeking behavior. These effects are visualized in figure 15. Confirmation of these expectations can greatly contribute to our knowledge considering neurobiological circuits underlying loss of control and compulsivity as important aspects of alcohol use disorder. In summary the present study provides essential evidence that the conditioned suppression model can be used to model individual differences in alcohol consumption and loss of control over alcohol use. The model has an excellent face validity combining voluntary alcohol consumption with an excellent punishment paradigm for accurately modelling loss of control. In the second experiment the described model will be combined with an optogenetic approach to examine the effects of the prelimbic area and the dorsal lateral striatum in alcohol seeking behavior. The results of the second experiment can greatly increase our knowledge of the ne
urobiological circuits underlying loss of control and compulsivity as important aspects of alcohol use disorders. The combined approach of the conditioned suppression model and optogenetics will lead to new promising targets for future research and treatment considering people suffering from AUDs. This study demonstrates that the conditioned suppression model is the best way to accurately examine loss of control, an essential aspect in the development of an AUD. Using this model we can examine the possibilities of implementing regaining control over alcohol use in future clinical treatment for AUD. References 1. Global status report on alcohol and health. . . http://www.who.int/substance_abuse/publications/global_alcohol_report/msbgsruprofiles.pdf.
2. American Psychiatric Association. American psychiatric association: Diagnostic and statistical manual of mental disorders. Fifth Edition ed. Arlington: American Psychiatric Association; 2013.
3. Nutt DJ, King LA, Phillips LD. Drug harms in the UK: A multicriteria decision analysis. The Lancet. 2010;376(9752):1558-1565.
4. Wackernah RC, Minnick MJ, Clapp P. Alcohol use disorder: Pathophysiology, effects, and pharmacologic options for treatment. Substance abuse and rehabilitation. 2014;5:1.
5. Miller WR, Wilbourne PL. Mesa grande: A methodological analysis of clinical trials of treatments for alcohol use disorders. Addiction. 2002;97(3):265-277.
6. Vanderschuren LJ, Ahmed SH. Animal studies of addictive behavior. Cold Spring Harb Perspect Med. 2013;3(4):a011932. doi: 10.1101/cshperspect.a011932 [doi].
7. Hopf FW, Lesscher HM. Rodent models for compulsive alcohol intake. Alcohol. 2014;48(3):253-264.
8. Lesscher H, Vanderschuren LJ. Compulsive drug use and its neural substrates. . 2012.
9. Substance-related and addictive disorders. In: American Psychiatric Association; 2013. http://dx.doi.org/10.1176/appi.books.9780890425596.dsm16. doi:10.1176/appi.books.9780890425596.dsm16.
10. Vanderschuren LJ, Everitt BJ. Drug seeking becomes compulsive after prolonged cocaine self-administration. Science. 2004;305(5686):1017-1019. doi: 10.1126/science.1098975 [doi].
11. Limpens JH, Schut EH, Voorn P, Vanderschuren LJ. Using conditioned suppression to investigate compulsive drug seeking in rats. Drug Alcohol Depend. 2014;142:314-324.
12. Lesscher H, Van Kerkhof LW, Vanderschuren LJ. Inflexible and indifferent alcohol drinking in male mice. Alcoholism: Clinical and Experimental Research. 2010;34(7):1219-1225.
13. Wolffgramm J. An ethopharmacological approach to the development of drug addiction. Neuroscience & Biobehavioral Reviews. 1992;15(4):515-519.
14. Hopf FW, Chang S, Sparta DR, Bowers MS, Bonci A. Motivation for alcohol becomes resistant to quinine adulteration after 3 to 4 months of intermittent alcohol Self’Administration. Alcoholism: Clinical and Experimental Research. 2010;34(9):1565-1573.
15. Deroche-Gamonet V, Belin D, Piazza PV. Evidence for addiction-like behavior in the rat. Science. 2004;305(5686):1014-1017. doi: 10.1126/science.1099020 [doi].
16. Wolffgramm J, Heyne A. From controlled drug intake to loss of control: The irreversible development of drug addiction in the rat. Behav Brain Res. 1995;70(1):77-94.
17. M. Spoelder. Individual differences in quinine adulteration in alcohol drinking rats. . 2015.
18. Lesscher et al. Conditioned suppression: Limited versus extended alcohol exposure. . 2015.
19. Spoelder, M., Vanderschuren, L.J.M.J. & Lesscher, H.M.B. Escalation of alcohol intake in lister hooded rats on an intermittent every-other-day schedule of alcohol consumption. . 2013.
20. Voges J, Mï¿½ï¿½ller U, Bogerts B, Mï¿½ï¿½nte T, Heinze H. Deep brain stimulation surgery for alcohol addiction. World Neurosurgery. 2013;80(3’4):S28.e21-S28.e31. doi: http://dx.doi.org/10.1016/j.wneu.2012.07.011.
21. Mishra BR, Nizamie SH, Das B, Praharaj SK. Efficacy of repetitive transcranial magnetic stimulation in alcohol dependence: A sham’controlled study. Addiction. 2010;105(1):49-55.
22. Everitt BJ, Robbins TW. Neural systems of reinforcement for drug addiction: From actions to habits to compulsion. Nat Neurosci. 2005;8(11):1481-1489.
23. Chen BT, Yau H, Hatch C, et al. Rescuing cocaine-induced prefrontal cortex hypoactivity prevents compulsive cocaine seeking. Nature. 2013;496(7445):359-362.
24. Corbit LH, Nie H, Janak PH. Habitual alcohol seeking: Time course and the contribution of subregions of the dorsal striatum. Biol Psychiatry. 2012;72(5):389-395.
25. Everitt BJ, Robbins TW. Neural systems of reinforcement for drug addiction: From actions to habits to compulsion. Nat Neurosci. 2005;8(11):1481-1489.
26. Murray JE, Belin D, Everitt BJ. Double dissociation of the dorsomedial and dorsolateral striatal control over the acquisition and performance of cocaine seeking. Neuropsychopharmacology. 2012;37(11):2456-2466.
27. Dickinson A, Wood N, Smith JW. Alcohol seeking by rats: Action or habit? The Quarterly Journal of Experimental Psychology: Section B. 2002;55(4):331-348.
28. Jonkman S, Pelloux Y, Everitt BJ. Differential roles of the dorsolateral and midlateral striatum in punished cocaine seeking. J Neurosci. 2012;32(13):4645-4650. doi: 10.1523/JNEUROSCI.0348-12.2012 [doi].
29. Belin D, Everitt BJ. Cocaine seeking habits depend upon dopamine-dependent serial connectivity linking the ventral with the dorsal striatum. Neuron. 2008;57(3):432-441.
30. Vanderschuren LJ, Di Ciano P, Everitt BJ. Involvement of the dorsal striatum in cue-controlled cocaine seeking. J Neurosci. 2005;25(38):8665-8670. doi: 25/38/8665 [pii].
31. Limpens JHW, Damsteegt R, Broekhoven MH, Voorn P, Vanderschuren LJMJ. Pharmacological inactivation of the prelimbic cortex emulates compulsive reward seeking in rats. Brain Res. (0). doi: http://dx.doi.org/10.1016/j.brainres.2014.10.045.
32. Tye KM, Deisseroth K. Optogenetic investigation of neural circuits underlying brain disease in animal models. Nature Reviews Neuroscience. 2012;13(4):251-266.
33. Airan RD, Thompson KR, Fenno LE, Bernstein H, Deisseroth K. Temporally precise in vivo control of intracellular signalling. Nature. 2009;458(7241):1025-1029.
34. Vanderschuren LJMJ. Shining light on loss of control over sustance and food intake. . 2014.
35. Simms JA, Steensland P, Medina B, et al. Intermittent access to 20% ethanol induces high ethanol consumption in Long’Evans and wistar rats. Alcoholism: Clinical and Experimental Research. 2008;32(10):1816-1823.
36. Wise RA. Voluntary ethanol intake in rats following exposure to ethanol on various schedules. Psychopharmacologia. 1973;29(3):203-210.
37. Sparta DR, Stamatakis AM, Phillips JL, Hovelsï¿½ï¿½ N, van Zessen R, Stuber GD. Construction of implantable optical fibers for long-term optogenetic manipulation of neural circuits. Nature protocols. 2012;7(1):12-23.
38. Vendruscolo LF, Roberts AJ. Operant alcohol self-administration in dependent rats: Focus on the vapor model. Alcohol. 2014;48(3):277-286.
39. Lieber CS, DeCarli LM. Liquid diet technique of ethanol administration: 1989 update. Alcohol Alcohol. 1989;24(3):197-211.
40. Carnicella S, Ron D, Barak S. Intermittent ethanol access schedule in rats as a preclinical model of alcohol abuse. Alcohol. 2014;48(3):243-252.
41. Seif T, Chang S, Simms JA, et al. Cortical activation of accumbens hyperpolarization-active NMDARs mediates aversion-resistant alcohol intake. Nat Neurosci. 2013;16(8):1094-1100.
42. Pelloux Y, Everitt BJ, Dickinson A. Compulsive drug seeking by rats under punishment: Effects of drug taking history. Psychopharmacology (Berl ). 2007;194(1):127-137.
In this study we also focused on adolescent drinking behavior and the effects on drinking behavior in adulthood. Unlike the data of the conditioned suppression the presented data is not combined with data from another study. The effects of adolescent drinking behavior and the predictive value of adolescent drinking for drinking behavior in adulthood or AUD development in adulthood are still unknown. Preclinical research can help us to gain more knowledge considering the effects of drinking in adolescence on drinking behavior in adulthood. This study looked in Lister Hooded rats (n=48) at the effects of alcohol exposure during adolescence and the effects on alcohol drinking behavior of the same animals in adulthood. Two groups were compared; one group which received alcohol during adolescence for two weeks(n=24, AdolAlc) and one group which did not receive alcohol during adolescence for two weeks (n=24, AdolWat). In adulthood groups were exposed to alcohol following the adapted intermittent-every-other day schedule and afterwards animals were classified in subgroups following their alcohol intake. The results considering the 24H sessions are displayed in supplementary figure 1. The most important effect is that of adolescent alcohol consumption on adulthood alcohol intake. The adolescent alcohol group drinks significantly more than the adolescent water group in adulthood (Fadolescence(1,42)= 213,379, p<0,01) and there is a significant effect between the subgroups(Fsubgroup(2,42) = 130,812, p<0,01). This shows that alcohol exposure during adolescence predicts a higher average level of alcohol intake in adulthood in rats. These results indicate that adolsescent drinking should be considered as a possible contributing factor for escalation of alcohol use in adulthood. Supplementary figure 1. This figure displays the effects of drinking in adolescence within each drinking subgroup. The group which received alcohol during adolescence (Adol EtOH)has a significant higher average EtOH intake per 24H session than the group which did not receive alcohol during adolescence (Adol Water) in all the different drinking subgroups (low,medium,high)(p<0,01). Additively drinking subgroups(low vs. medium vs. high) also differed significantly from each other(p<0,01).