Examination of the effects of cannabinoid ligands on decision making in a rat gambling task

Accepted Manuscript Examination of the effects of cannabinoid ligands on decision making in a rat gambling task

Jacqueline-Marie N. Ferland, Madison Carr, Angela M. Lee, Myrthe E. Hoogeland, Catharine A. Winstanley, Tommy Pattij PII: DOI: Reference:

S0091-3057(17)30705-0 doi:10.1016/j.pbb.2018.05.012 PBB 72604

To appear in:

Pharmacology, Biochemistry and Behavior

Received date: Revised date: Accepted date:

18 December 2017 8 May 2018 18 May 2018

Please cite this article as: Jacqueline-Marie N. Ferland, Madison Carr, Angela M. Lee, Myrthe E. Hoogeland, Catharine A. Winstanley, Tommy Pattij , Examination of the effects of cannabinoid ligands on decision making in a rat gambling task. The address for the corresponding author was captured as affiliation for all authors. Please check if appropriate. Pbb(2017), doi:10.1016/j.pbb.2018.05.012

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT Examination of the effects of cannabinoid ligands on decision making in a rat gambling task

Jacqueline-Marie N. Ferland1,2,3, Madison Carr1, Angela M. Lee1,4, Myrthe E. Hoogeland1,

RI

Department of Anatomy and Neurosciences, Amsterdam Neuroscience, VU University

SC

1

PT

Catharine A. Winstanley2, Tommy Pattij1*

Medical Center, Amsterdam, the Netherlands

Department of Psychology, Djavad Mowafaghian Centre for Brain Health, University of

NU

2

British Columbia, Vancouver, Canada

Current Address: Department of Psychiatry, Mount Sinai School of Medicine, New York, NY,

MA

3

USA

D

Current address: Department of Psychiatry, Yale School of Medicine, New Haven, CT, USA

T. Pattij, Ph.D.

CE

*Corresponding author:

PT E

4

AC

Department of Anatomy and Neurosciences, Amsterdam Neuroscience VU University Medical Center De Boelelaan 1108, 1081 HZ Amsterdam, The Netherlands Phone: +31 20 44 48040; Fax: +31 20 44 48100 E-mail: [email protected]

1

ACCEPTED MANUSCRIPT Abstract Although exposure to delta-9-tetrahydrocannabinol (THC) is perceived to be relatively harmless, mounting evidence has begun to show that it is associated with a variety of cognitive deficits, including poor decision making. THC-induced impairments in decision

PT

making are thought to be the result of cannabinoid CB1 receptor activation, and although clinical literature suggests that chronic activation via THC contributes to perturbations in

RI

decision making, acute CB1 receptor modulation has yielded mixed results. Using an animal

SC

model to examine how CB1-specific ligands impact choice biases would provide significant

NU

insight as to how recruitment of the endocannabinoid system may influence decision making. Here, we used the rat gambling task (rGT), a validated analogue of the human Iowa

MA

Gambling Task, to assess baseline decision making preferences in male Wistar rats. After acquisition rGT performance was measured. Animals were challenged with the CB1 receptor

D

antagonist rimonabant, the partial agonist THC, and the synthetic agonist WIN 55,212-2.

PT E

Animals were also treated acutely with the fatty acid amide hydrolase (FAAH) inhibitor URB597 to selectively upregulate the endocannabinoid anandamide. Blockade of the CB1

CE

receptor produced a trend improvement in decision making in animals who preferred the advantageous task options, yet left choice unaffected in risk-prone rats. Neither CB1

AC

receptor agonist had strong effects on decision making, but a high dose THC decreased premature responses, whereas WIN55,212-2 did the opposite. URB597 did not affect task performance. These results indicate that although chronic CB1 receptor activation may be associated with impaired decision making, acute modulation has modest effects on choice and instead may play a substantive role in regulating impulsive responding.

Keywords: cannabinoid; cognition; decision making; gambling; rat 2

ACCEPTED MANUSCRIPT 1. Introduction Legislation permitting cannabis use is on the rise in several western countries, leading to greater consumption (United Nations, 2015). Use of the active ingredient in cannabis, Δ9tetrahydrocannabinol (THC), is widely viewed to be relatively harmless and having little to no

PT

addiction potential (United Nations, 2015). However, despite social acceptance, recent studies portray a different picture of THC, with use being associated with increased

RI

prevalence of a variety of psychiatric phenotypes including addiction, depression, anxiety,

SC

and schizophrenia (Renard et al., 2014). Interestingly, poor cost/benefit decision making has

NU

been implicated in each of these conditions (Bechara, 2005; Cella et al., 2010; Zhang et al., 2015; Pedersen et al., 2017), suggesting deficits in choice may serve as a common

MA

contributor to expression of these psychopathologies, and may be affected by THC consumption. Indeed, data from studies using the Iowa Gambling Task (IGT), a clinical

D

paradigm designed to operationalize cost/benefit decision making, show chronic cannabis

PT E

use is also associated with maladaptive choice (Bolla et al., 2005). Therefore, there may be overlapping circuitries that are affected by THC which culminates in psychiatric illness and

CE

poor decision making. Investigating the role of cannabinoid CB1 receptor activity, the main target of THC, on decision making may provide unique insight into the etiology of these

AC

disorders, and contribute to development of treatments. Though chronic cannabis experience is associated with impaired decision making, it is unclear what aspect of decision making is the most affected, such as motivation or value processing, and whether effects on choice are present even after acute THC experience. One human study found acute THC exposure reduced effort-based decision making, especially when probabilities of winning were low, suggesting THC fosters an a-motivational state (Lawn et al., 2016). This data is also supported by animal studies (Silveira et al., 2017), 3

ACCEPTED MANUSCRIPT where THC dose-dependently reduced willingness to expend cognitive effort. Another study found acute THC administration in occasional users produced impairments in reaction time on a monetary incentive delay task, but was also associated with greater orbitofrontal gyrus activation during reward anticipation (Van Hell et al., 2012), indicating THC may impact sensitivity to reward. However, results from human reports are mixed on the recruitment of

PT

the endocannabinoid system and probabilistic decision making; results from abstinent

RI

cannabis users with a history of heavy intake showed impaired decision making on the IGT,

SC

but other studies found no impairments after acute THC administration or in active heavy users (Bolla et al., 2005; Ramaekers et al., 2006; Cousijn et al., 2013). These data suggest

NU

cognitive impairment after THC experience may be due to length of use rather than general

MA

deleterious effects of the drug. However, another explanation is that the variable effects seen in humans may be due to individual differences in activation of brain regions important

D

for reward evaluation and decision making (Cousijn et al., 2013), suggesting that differences

PT E

in reward processing may explain discrepancies in responding to THC, and potentially even demonstrate a vulnerability to cognitive deficits after acute THC exposure.

CE

A significant challenge when evaluating human data are the myriad of social, cultural, and environmental factors which can influence results. Animal models provide a means to

AC

investigate the causal relationship between behavior and its neurobiological substrates (Potenza, 2009). In recent years, using a rodent analogue of the IGT, the rat gambling task (rGT), we have shown animals produce decision making profiles akin to humans, and have explored the influence of several neurotransmitter systems on risky decision making (Zeeb et al., 2009; De Visser et al., 2011; Winstanley and Clark, 2016). Furthermore, the task measures additional variables, such as impulse control and motivation, giving the paradigm greater behavioral resolution and, making it ideally suited to study how different 4

ACCEPTED MANUSCRIPT neurobiological mechanisms contribute to distinct cognitive processes. Indeed, we have found previously that serotonin and dopamine modulate different aspects of rGT performance (Zeeb et al., 2009). Interestingly, activation of the cannabinoid CB1 receptor mediates dopaminergic output (French, 1997) and dorsal raphe activity (Haj-Dahmane and Shen, 2009), influences responding to food and drug rewards (Sanchis-Segura et al., 2004; De

PT

Vries et al., 2005), and affects the maturation of key regions implicated in decision making,

RI

including the prefrontal cortex and amygdala, during development (Parsons and Hurd, 2015;

SC

Zeeb and Winstanley, 2011, 2013). Therefore, THC may influence decision making by mediating dopaminergic and serotonergic mechanisms, although cannabinoid-mediated

NU

effects on glutamate and GABA transmission could also impact decision making (Colizzi et al.,

MA

2016; Piomelli, 2003). Another powerful asset of the rGT is the ability to assess individual differences in decision making. There is an emerging body of preclinical work that has shown

D

baseline patterns of risky choice can influence responding to drugs of abuse (Ferland and

PT E

Winstanley, 2017; Mitchell et al., 2014). It is possible that the variable effects seen after THC administration may be representative of differential responding to rewards, and animals

CE

prone to select large, risky rewards may be more sensitive to CB1 activation. Using the rGT, we can assess how animals respond to acute administration of THC and other CB1 agents,

humans.

AC

and perhaps clarify why drugs like THC and synthetic cannabinoids produce mixed effects in

To explore the role of the CB1 receptor in decision making, this study employed acute administration of CB1 receptor agonists THC and WIN 55,212-2, the CB1 receptor antagonist rimonabant, and the fatty acid amide hydrolase (FAAH) inhibitor URB-597 to determine the influence of cannabinoid CB1 receptor activity on choice preferences, impulsive action, and motivation in male Wistar rats. We also examined the potential role of individual risk 5

ACCEPTED MANUSCRIPT preference in the response to these drugs. To replicate previous findings obtained in other rat strains, we administered acute doses of the psychostimulant amphetamine. We hypothesized that administration of CB1 receptor agonists would impair, while CB1 antagonism would improve, task performance, and individual sensitivity to risk would

AC

CE

PT E

D

MA

NU

SC

RI

PT

mediate the behavioral outcome to CB1 modulation.

6

ACCEPTED MANUSCRIPT 2. Methods

2.1. Subjects In total thirty-two male Wistar rats were obtained from Harlan CPB (Horst, The

and

were

housed

two

per

cage

in

macrolon

PT

Netherlands). At the start of the experiments animals weighed approximately 250 grams, cages

(42.526.618.5

cm;

RI

lengthwidthheight) under a reversed 12 h light/dark cycle (lights on at 7.00 PM) at

SC

controlled room temperature (21 ± 2 ºC) and relative humidity (60 ± 15%). Animals were

NU

maintained at approximately 90% of their free-feeding weight, starting one week prior to the start of experiments by restricting the amount of standard rodent food chow. Water was

MA

available ad libitum throughout the entire experiment. All experiments were conducted with the approval of the animal ethical committee of the VU University and VU University Medical

PT E

D

Center Amsterdam, the Netherlands, and all efforts were made to minimize animal suffering.

2.2. Apparatus

CE

Experiments were conducted in sixteen identical rat five-hole nose poke operant

AC

chambers with stainless steel grid floors (MED-NPW-5L, Med Associates Inc., Fairfax, VT, USA) housed in sound-insulating and ventilated cubicles. Set in the curved wall of each box was an array of five circular holes 2.54 cm in diameter, 2.2 cm deep and 2.25 cm above floor level. Each hole was equipped with an infrared detector located across each nose poke unit 1.0 cm from the front, and a yellow LED stimulus light (6.4 mm in diameter). Rodent Dustless Precision Pellets® (45 mg, F0021, Bio-Serv, Frenchtown, NJ, USA) were delivered at the opposite wall via a dispenser into a food tray. In addition, the chamber could be illuminated by a white house light. A computer equipped with MED-PC version 1.17 (Med Associates Inc., 7

ACCEPTED MANUSCRIPT Fairfax, VT, USA) controlled experimental sessions and recorded data. Animals were trained once daily from Monday to Friday, during the dark phase of the light/dark cycle.

2.3. Behavioral procedure Two cohorts of animals (n = 16 each) were trained in the rGT as described below.

PT

Upon establishment of a stable baseline, animals in cohort 1 were pharmacologically

RI

challenged with the cannabinoid CB1 receptor antagonist rimonabant, the synthetic

SC

cannabinoid CB1 receptor agonist WIN55,212-2, and the cannabinoid CB1 receptor agonist Δ9-tetrahydrocannabinol (THC). In cohort 2, the effects of the FAAH inhibitor URB597 and

NU

the psychostimulant amphetamine were examined. For both cohorts, identical training

MA

protocols were followed.

D

2.3.1. Habituation

PT E

Animals were habituated to food pellet rewards and the operant chamber over 7 days. On days 1-3, ten food pellets were placed in each home cage. Food restriction began

CE

on day 3. On days 4 and 5, animals were placed in the operant chamber for 20 min each day with the house light on and five food pellets in the food tray. On days 6 and 7, animals were

AC

placed in the operant chamber and 75 pellets were dispensed on average every 15 s, so that animals would associate the sound of pellet delivery with reward.

2.3.2. rGT Training After habituation as described above, animals were subjected to a series of rGTspecific training protocols, described in detail in previously published reports (Zeeb et al., 2009; Zeeb and Winstanley, 2011). As previously described, the rGT mimics the human Iowa 8

ACCEPTED MANUSCRIPT Gambling Task by presenting four choices with different reward and punishment amounts and probabilities (Bechara et al., 1994; Zeeb et al., 2009). First, the animals were trained to make a nose-poke response in an illuminated hole in order to earn a reward in the food tray; the illuminated hole varied between trials across holes 1, 2, 4, and 5. Each session consisted of 100 trials and lasted for approximately 30 min. This phase of training involved three

PT

stages, with stimulus duration (i.e. light on in the nose-poke hole), limited hold (i.e. delay

RI

between stimulus light extinction and trial end), and the inter-trial interval programmed for

SC

30 s, 30 s, and 2 s, respectively, in stage 1, then 20 s, 20 s, and 2 s in stage 2, and 10 s, 10 s, and 5 s in stage 3. Correct responses were those made at the illuminated hole. Omissions

NU

were counted when no response was made during stimulus illumination or the limited hold.

MA

The criteria for progressing to the next stage were completion of 100 trials with ≥80% correct and ≤20% omitted. Animals not meeting these criteria after three sessions of stage 1

D

training were placed on an additional training protocol to ensure that they learned to

PT E

respond at illuminated nose-poke holes to earn reward; holes 1, 2, 4 and 5 were illuminated all at once, and then individually, with more than 75 responses required to progress from

CE

each stage. All animals were meeting criteria in stage 3 of rGT training after 10 days. Second, animals were trained to associate each nose-poke hole with the reward and

AC

punishment contingencies it would maintain throughout the remainder of the experiment. This training was accomplished through 7 sessions in a forced choice version of the rGT in which all holes were programmed with their respective reward amount, punishment amount, and punishment probability (Table 1), but only one nose poke hole was lit in each trial as opposed to the full rGT which presents the four nose poke holes lit together. The lit option varied between trials in pseudorandom order. All other functions were executed as in the full rGT described below. Beginning with the forced choice rGT, animals were divided 9

ACCEPTED MANUSCRIPT into two equal groups, with cage mates in the same group, to counterbalance choice and nose-poke hole pairings outlined as version A and version B. Finally, animals were given the full rGT, which recorded their choice behavior, premature responses (a measure of motor impulsivity), omissions, number of rewards earned, percent of trials rewarded, perseverative responses, and latencies to choose and

PT

reward collection (proxies of choice deliberation and motivation to consume reward). The 30

RI

min session, as well as each trial within the session, was initiated by a nose-poke in the lit

SC

food tray. After a 5 s inter-trial interval with the tray-light off, holes 1, 2, 4, and 5 were lit simultaneously for 10 s, during which the animals could make a response choice. If a

NU

response was made into nose-poke holes 1, 2, 4, or 5 during the inter-trial interval, it was

MA

recorded as a premature response (a measure of motor impulsivity) and punished by a 5 s time-out period. A response during the 10 s stimulus duration was either rewarded or

D

punished according to the reinforcement schedule associated with that choice, and the

PT E

choice was recorded. If rewarded, the proper number of pellets was delivered to the food tray and a subsequent response at the food tray would initiate the next trial. If punished, all

CE

lights were extinguished except the light in the chosen hole, which flashed at 0.5 Hz for the duration of the time-out. Failure to respond within 10 s after stimulus illumination was

AC

measured as an omitted response (a measure of motivation), and after which another trial would begin.

The following measures for decision making were calculated at baseline and during drug tests: 1) percent choice = (number of choices made for a pellet option)/(total number of choices or responses made) x 100; 2) choice score, which is a single value calculated from the percent choice of the two advantageous options (P1+P2) minus the percent choice of the two disadvantageous options (P3+P4). Animals with a positive choice score were designated 10

ACCEPTED MANUSCRIPT as ‘optimal’, whereas animals with a negative choice score were designated as ‘riskpreferring’. In addition, other measures analyzed were the percentage premature responses, number of omissions, number of trials completed and choice latencies and reward collection latencies.

PT

2.4. Drugs

RI

Rimonabant (SR141716A; 0.3, 1.0 and 3.0 mg/kg) and THC (0.5, 1.0 and 2.0 mg/kg)

SC

were generated, and kindly provided by, AbbVie and Echo Pharmaceuticals (both Weesp, the Netherlands), respectively. (+)-Amphetamine sulfate (OPG, Utrecht, the Netherlands; 1.0

NU

mg/kg) was dissolved in sterile saline, whereas rimonabant, URB597 (Sigma Aldrich, St. Louis,

MA

MO, USA; 0.3 mg/kg), WIN55,212-2 (Tocris Bioscience, Avonmouth, UK; 0.3, 1.0 and 3.0 mg/kg), and THC were dissolved in a mixture of 96% ethanol, Tween80, and sterile saline

D

(ratio 1:1:18) as described before (De Vries et al., 2001). Drug doses and injection times were

PT E

based on previous studies (Pattij et al., 2007a, b; Wiskerke et al., 2011b): URB597 was injected 120 min prior to testing, rimonabant 45 min prior to testing, THC 30 min prior to

CE

testing, and WIN55,212-2 and amphetamine were injected 20 min before testing. Drugs were freshly prepared on each test day and injected intraperitoneally (i.p.) in a volume of 1

AC

mL/kg bodyweight. Rimonabant, THC and WIN55,212-2 were administered according to a Latin square within-subjects design, whereas for amphetamine and URB597 half of the animals received a drug dose on the first test and vehicle on the second test day. For the other half of the animals this order was reversed. Drug challenges were conducted on Tuesdays and Fridays with baseline training sessions on the other weekdays. Prior to the first test day, all animals had been habituated twice to saline injections. Tests with the different

11

ACCEPTED MANUSCRIPT compounds were separated by a washout period of one week, during which the animals were trained in the rGT without drug challenges.

2.5. Data analysis Data were analyzed using IBM SPSS Statistics version 20 (IBM Corporation, Armonk,

PT

NY, USA) and were subjected to repeated measures analysis of variance (ANOVA) with drug

RI

dose and choice as within subjects variables and choice preference (‘optimal’ vs ‘risk-

SC

preferring’) as a between subjects variable. All data expressed as percentages (i.e. choice and premature responding) were arcsine transformed before analyses to limit the effect of

NU

an artificially-imposed ceiling (McDonald, 2009). The homogeneity of variance across groups

MA

was determined using Mauchly’s tests for equal variances and in case of violation of homogeneity, Huynh-Feldt epsilon (ε) adjusted degrees of freedom resulting in more

D

conservative probability values were depicted and used for subsequent analyses. In cases of

PT E

statistically significant main dose effects, further paired comparisons were conducted using paired t-tests. The level of probability for statistically significant effects was set at P = 0.05.

CE

Graphs were produced using GraphPad Prism version 5.02 for Windows (GraphPad Software,

AC

San Diego, CA, USA).

12

ACCEPTED MANUSCRIPT 3. Results 3.1. Baseline performance Selection of animals based on their choice score during stable baseline performance yielded in total N = 20 optimal rats and N = 12 risk-preferring rats (cohort 1: n = 10 optimal

PT

and n = 6 risk-preferring rats; cohort 2: n = 10 optimal and n = 6 risk-preferring rats). Under stable baseline performance across 3 sessions, optimal animals showed a significant

RI

preference for P2 over the other pellet choice options and risk-preferring rats for P2, P3 and

SC

P4 over P1 (Fig. 1A; F(3,84) choice = 11.72, P < 0.001; F(3,84) choice x group = 7.72, P < 0.0001;

NU

F(1,28) group = 2.16, NS; F(2,56) session = 0.13, NS; F(6,168) session x choice x group = 1.03, NS). This group difference was also reflected in the animals’ choice score across baseline sessions (Fig.

MA

1B; F(1,28) group = 60.26, P < 0.0001; F(2,56) session = 0.003, NS; F(2,56) session x cohort x group = 0.90, NS), with optimal animals showing significantly higher choice scores. As indicated in Table 2,

session

= 0.43, NS; F(2,56)

PT E

< 0.0001; F(2,56)

D

risk-preferring rats also completed significantly fewer trials at baseline (F(1,28) group = 7.47, P session x cohort x group

= 0.47, NS), likely due to the

lengthy timeouts associated with the choice of the disadvantageous options of the task and

F(2,56)

group

= 17.77, P < 0.0001; F(2,56)

session

= 1.04, NS;

CE

earned significantly less pellets (F(1,28) session x cohort x group

= 0.42, NS). There were neither significant group differences in

AC

premature responding, omissions, and reward collection latencies, nor were there main session or main session x cohort x group interaction effects for these three variables (all F’s < 2.05, NS). Finally, while groups did not differ in terms of choice latencies, there was a small and significant change in choice latencies over the three baseline sessions (F(1,28) 0.002, NS; F(2,56)

session

= 8.45, P < 0.01; F(2,56)

session x cohort x group

group

=

= 2.05, NS). Further

comparisons revealed that collection latencies were slower in the third baseline session compared to both the first and second baseline session (baseline session 1: 1.20 ± 0.09 s; 13

ACCEPTED MANUSCRIPT baseline session 2: 1.17 ± 0.08 s; baseline session 3: 1.30 ± 0.1 s; P < 0.05 and P < 0.0001, respectively). During the drug testing phase, within-cohort analyses indicated no significant changes in decision making under vehicle conditions across drug challenges (cohort 1: F(6,84) test x choice x group = 0.46, NS ; cohort 2: F(3,42) test x choice x group = 0.64, NS), suggesting that

RI

PT

there was no drift in baseline behavior across drug tests.

SC

3.2. Effects of rimonabant on rGT performance

NU

Choice

choice

dose

= 2.19, NS; F(3,42)

MA

Rimonabant largely left decision making unaffected (F(3,42)

= 3.66, P < 0.05; F(3,42) choice x group = 5.62, P < 0.01; F(5,63) dose x choice = 0.41, NS; F(5,63) = 1.36, NS; F(1,14) group = 0.47, NS). However, choice was trending towards a

D

dose x choice x group

PT E

differential effect between the two groups, most notably in optimal animals visibly showing greater advantageous choice (Fig. 2A,B; F(3,42) dose x group = 2.70, P = 0.06), a result supported

0.08; F(3,42)

dose

CE

by a trend towards improved choice score in these rats (Fig. 2C; F(3,42) dose x group = 2.39, P = = 0.60, NS; F(1,14)

group

= 13.81, P < 0.0001). In contrast, risk-preferring

AC

animals tended to show the opposite effect.

Auxiliary measures Administration of rimonabant did not affect motor impulsivity (Fig. 2D; F(3,42) dose = 1.80, NS; F(3,42) dose x group = 0.37, NS; F(1,14) group = 0.39, NS), but did increase choice latencies in all animals (Table 4; F(3,42) dose = 5.79, P < 0.01; F(3,42) dose x group = 1.62, NS; F(1,14) group = 0.33, NS). This effect was greatest at the highest dose (vehicle vs 3 mg/kg: t(15) = -4.22, P < 14

ACCEPTED MANUSCRIPT 0.01; other doses: t(15) < 1.78, NS). Rimonabant did not affect collection latencies (dose and dose x group Fs < 2.92, NS), nor the number of trials completed; there was only a main effect of group, with risk-preferring rats continuing to complete fewer trials (Table 3; F(1,14) group = 14.43, P < 0.01; dose and dose x group Fs < 2.51, NS). Finally, administration of rimonabant

PT

did not alter the number of omissions (Table 3; all Fs < 1.63, NS).

RI

3.3. Effects of THC and WIN 55,212-2 on rGT performance

SC

Choice

The CB1 receptor agonists THC and WIN 55,212-2 had no significant impact on

0.03, NS; F(4,49) dose x choice = 0.84, NS; F(4,49) dose x choice x group = 1.36, NS; F(1,11) group

MA

dose x group =

NU

decision making at the individual choice level (Fig. 3A,B; THC: F(3,33) dose = 0.31, NS; F(3,33)

= 0.56, NS; Fig. 3E,F; WIN55,212-2: F(3,21) dose = 1.61, NS; F(3,21) dose x group = 0.52, NS; F(9,63) = 1.67, NS; F(9,63) dose x choice x group = 1.65, NS; F(1,7) group = 1.77, NS). Choice scores

D

dose x choice

PT E

trended to be differentially affected by THC but not WIN55,212-2, where risk-preferring rats appeared to have a better choice score after the highest dose (Fig. 3C; THC: F(2,21) dose =

CE

0.23, NS; F(3,39) dose x group = 2.57, P = 0.07; F(1,13) group = 17.54, P < 0.001; Fig. 3G; WIN:

AC

F(3,21) dose = 2.32, NS; F(3,21) dose x group = 2.35, NS; F(1,7) group = 14.56, P < 0.01).

Auxiliary measures

THC significantly reduced premature responding in all animals (Fig. 3D; F(3,42) dose = 5.70, P < 0.01; F(3,42) dose x group = 0.50, NS; F(1,14) group = 0.008, NS), an effect found to be significant at the highest dose (vehicle vs 0.5 mg/kg: t(15) = 1.25, NS; vehicle vs 1 mg/kg: t(15) = 0.85, NS; vehicle vs 2 mg/kg: t(15) = 4.03, P < 0.01). WIN 55,212-2 visibly affected premature responding in risk-preferring rats and optimal animals (Fig. 3H; F(3,24) dose = 0.95, 15

ACCEPTED MANUSCRIPT NS; F(3,24) dose x group = 0.91, NS; F(1,8) group = 9.24, P < 0.05). Post-hoc t-tests revealed the group difference in premature responding was significant at the middle and highest dose administered (t(14) veh = 0.57, NS; t(14) 0.5 mg/kg = -0.02, NS; t(14) 1 mg/kg = -1.03, P < 0.005; t(14) 2 mg/kg

= -3.11, P < 0.05), suggesting divergent effects of this drug were greatest at these

doses.

PT

WIN55,212-2 also increased choice latencies in all animals (Table 4; F(2,25) dose = 6.50,

RI

P < 0.01; F(3,42) dose x group = 2.32, NS; F(1,14) group = 2.52, NS) with the greatest effect at the

SC

highest dose (vehicle vs 0.3 mg/kg: t(15) = 0.72, NS; vehicle vs 1 mg/kg: t(15) = -1.93, P = 0.07; vehicle vs 3 mg/kg: t(15) =-3.61, P < 0.01). THC also increased choice latencies, but only

= 0.39, NS; optimal: F(2,14) dose = 6.44, P < 0.05; risk-preferring: F(3,15) dose = 0.01, NS),

MA

group

NU

in optimal animals (Table 5; F(2,24) dose = 2.81, NS; F(3,42) dose x group = 3.11, P < 0.05; F(1,14)

and exclusively at the highest dose (vehicle vs 2 mg/kg: t(9) = -3.35, P < 0.01; other doses: t<-

D

1.009, NS). Conversely, collection latencies were unaffected by any dose of either drug

PT E

(Tables 4,5; all dose and dose x group Fs < 2.92, NS). However, risk-preferring rats showed overall significantly faster collection latencies in the THC and WIN55,212-2 experiments

CE

(THC: F(1,14) group = 9.77, P < 0.01; WIN55,212-2: F(1,14) group = 51.39, P < 0.00001). WIN55,212-2 dose-dependently reduced the number of trials completed by all

AC

animals (Table 4; F(2,26) dose = 6.78, P < 0.01; F(3,42) dose x group = 1.40, NS; F(1,14) group = 14.51, P < 0.01) and increased the number of omissions (Table 4; F(1,17) dose = 4.34, P ≤ 0.05; F(1,17) dose x group

= 0.40, NS; F(1,14)

group

= 0.23, NS). These effects were significant after

administration of 1 and 3 mg/kg (completed trials, vehicle vs 0.3 mg/kg: t(15) = 0.75, NS; vehicle vs 1 mg/kg: t(15) = 2.83, P < 0.05; vehicle vs 3 mg/kg: t(15) = 3.46, P < 0.01; omissions, vehicle vs 0.3 mg/kg: t(15) = -1.00, NS; vehicle vs 1 mg/kg: t(15) = -2.19, P ≤ 0.05; vehicle vs 3 mg/kg: t(15) = -2.40, P < 0.05). In contrast, THC did not influence the number of 16

ACCEPTED MANUSCRIPT trials completed, but risk-preferring rats consistently completed fewer trials across THC doses compared to optimal rats (Table 5; THC: F(1,14) group = 10.67, P < 0.01; dose and dose x group Fs < 2.51, NS). THC did not affect omissions (Table 5; all Fs < 1.63, NS).

3.4. Effects of URB 597 on rGT performance

PT

The single dose of the FAAH inhibitor URB 597 had no effect on any task variable (Fig. 4,

12.15, P < 0.01; trials: F(1,14) group = 12.94, P < 0.01; collection latencies: F(1,14) group =

SC

group =

RI

Table 6; F(1,22) choice = 7.92, P < 0.01; F(3,42) choice x group = 3.72, P < 0.05; choice score: F(1,14)

4.15, P = 0.06; all other Fs < 1.93, NS), suggesting that acute modulation of endogenous

MA

NU

endocannabinoid transmission does not affect cost/benefit decision making in this rat strain.

3.5. Effects of amphetamine on rGT performance

D

Choice

NS; F(1,11)

dose x group

PT E

Amphetamine significantly impaired decision making performance (F(1,11) dose = 0.32, = 0.35, NS; F(3,33)

choice

= 6.45, P < 0.01; F(3,33)

choice x group

= 3.70, P <

CE

0.05; F(3,33) dose x choice = 4.23, P < 0.05; F(3,33) dose x choice x group = 3.16, P < 0.05; F(1,11) group = 2.28, NS). As decision making was differentially affected in optimal versus risk-preferring

AC

rats, subsequent analyses showed that in optimal animals amphetamine primarily increased disadvantageous choice (Fig. 5A; F(1,8) dose = 0.001, NS; F(3,24) choice = 9.68, P < 0.01; F(3,24) dose x choice =

3.38, P < 0.05) by increasing choice of P4 (t(9) P4 = -2.99, P < 0.05). Risk-preferring

rats’ decision making was also affected by 1.0 mg/kg amphetamine (Fig. 5A; F(1,3) dose = 0.25, NS; F(3,9) (t(3)

P2

choice

= 2.40, NS; F(3,9)

= -5.03, P < 0.05; t(3)

dose x choice

P3

= 9.48, P < 0.01), increasing choice of P2 and P3

= 3.98, P < 0.05). Choice score was not affected by

17

ACCEPTED MANUSCRIPT amphetamine (Fig. 5B: F(1,11) group = 21.16, P < 0.01; F(1,11) dose = 0.034, NS; F(1,11) dose x group = 0.012, NS).

Auxiliary measures Amphetamine did not affect premature responding (Fig. 5C; all Fs < 1.44, NS).

dose x group

= 1.47, NS; F(1,14)

RI

preferring’ rats completed fewer trials (Table 7; F(1,14)

PT

Additionally, the number of completed trials was not affected by amphetamine, yet ‘riskgroup

=

SC

9.56, P < 0.01), likely due to the greater time spent in timeout as a result of choosing P4 more frequently.

= 5.57, P < 0.05), amphetamine increased optimal animals’ latency to collect reward

MA

group

NU

Though the group difference was maintained for reward collection latencies (Table 7; F(1,13)

(F(1,13) dose = 6.74, P < 0.05; F(1,13) dose x group = 34.53, P < 0.0001; optimal: F(1,8) dose = 45.42,

D

P < 0.001; risk-preferring: F(1,5) dose = 4.39, NS). Amphetamine had no effect on the number

PT E

of omissions in either group of rats (all Fs < 2.29, NS).

CE

4. Discussion

The current findings indicate that risky decision making as measured in the rGT can

AC

be reliably assessed in male Wistar rats, thereby extending the utility of this model to a previously untested rat strain. Moreover, consistent with previous data, pharmacological challenges with amphetamine affected decision making and promoted preference for disadvantageous choice, particularly in animals that displayed optimal decision making. In contrast, acute administration of cannabinoid ligands did not have strong effects on choice patterns in the rGT, yet did modulate motor impulsivity. In this regard, it should be pointed out that in both cohorts, the group size of ‘risk-preferring’ rats was relatively small (n=6). 18

ACCEPTED MANUSCRIPT Thus, it should be emphasized that in some aspects the current findings could have been underpowered to fully capture drug effects in this group of rats. Further establishing the acute effects of cannabinoids in the rGT would then require training larger cohorts of animals in the task to recruit higher sample sizes of ‘risk-preferring’ rats. The rGT used here was developed by Winstanley and colleagues (Zeeb et al., 2009)

PT

and the collective work using this model has clearly elucidated cholinergic, dopaminergic,

RI

noradrenergic and serotonergic modulation of risky decision making in various rat strains,

SC

including Long-Evans and Lister Hooded rats (Baarendse et al., 2013; Di Ciano et al., 2015; Silveira et al., 2015; Zeeb et al., 2009). Moreover, recent work in Long Evans rats has focused

NU

on individual differences in choice preference in this task by designating animals that prefer

MA

the more disadvantageous pellet options as ‘risk-preferring’ and animals that prefer the advantageous pellet options as ‘optimal’, and subsequently examining the interaction of

D

these baseline preferences with cocaine self-administration behavior (Ferland and

PT E

Winstanley, 2017). Since all of our previous work on cannabinoid modulation of executive functioning was conducted in Wistar rats (Pattij et al., 2007a; Wiskerke et al., 2011b, 2012),

CE

the current study employed male Wistar rats in the rGT. Similar to the other rat strains, Wistar rats readily achieved stable baseline performance in the task. In addition, we

AC

designated ‘risk-preferring’ rats and ‘optimal’ rats that stably differed in terms of their preference for the disadvantageous and advantageous response options. As an additional validation, acute challenges with 1.0 mg/kg amphetamine were used, and found to cause impaired decision making by promoting choice of the disadvantageous P4 pellet option in ‘optimal’ rats. This observation fits with previous data showing that amphetamine reduced preference of the advantageous P2 pellet option and increased preference for the P4 pellet option at comparable doses in Lister Hooded and Long-Evans rats (Baarendse et al., 2013; 19

ACCEPTED MANUSCRIPT Zeeb et al., 2009). Moreover, we found that in ‘risk-preferring’ rats there were modest, yet significant, increases in choice of both P2 and P3. Although further work is required, these differential responses towards amphetamine challenges might be caused by underlying differences in dopamine signaling between rats with an ‘optimal’ and ‘risk-preferring’ choice profile. In support of this notion, differences in striatal dopamine D2/D3 transmission have

PT

been linked to individual differences in rats’ preference for uncertainty in a translational

RI

betting task (Cocker et al., 2012).

SC

Surprisingly, amphetamine did not increase the number of premature responses in the rGT. Thus, in this respect Wistar rats differ from Lister Hooded and Long-Evans rats,

NU

which both display enhanced premature responding upon amphetamine challenges in this

MA

task (Baarendse et al., 2013; Zeeb et al., 2009). Whereas amphetamine did decrease response latencies and thus exerted clear psychomotor effects, possibly procedural

D

differences between the rGT and other instrumental tasks that assess premature responding

PT E

such as the 5-choice serial reaction time task (Bari et al., 2008) might explain amphetamine’s lack of effect in Wistar rats in the present study. In the 5-choice serial reaction time task

CE

employed in our laboratory, rats need to respond to a visual stimulus within 2 seconds after stimulus presentation and we have repeatedly found that amphetamine robustly increases

AC

premature responding (Van Gaalen et al., 2006; Pattij et al., 2007b; Wiskerke et al., 2011a,b). In the current rGT, rats withhold responding for 5 seconds prior to stimulus presentation, and have 10 seconds to make a choice between the different response options. Perhaps this difference in time given to respond could have contributed to the current observations of null effects of amphetamine on premature responding in Wistar rats. Previous data support this, and show amphetamine also failed to change impulsive or premature responding in

20

ACCEPTED MANUSCRIPT Wistar rats in an attention task with no requirements to inhibit prepotent responding (Bizot et al., 2015). Contrary to our hypotheses, we found no clear evidence supporting a role for the cannabinoid system in risk-based decisions. A recent study that used Long Evans rats showed a high dose of WIN55,212-2 promoted advantageous choice in animals making suboptimal

PT

choices only (Gueye et al., 2016). Somewhat in keeping with these data, acute THC

RI

challenges tended to promote advantageous choice in ‘risk-preferring’ rats (main dose x

SC

group interaction effect P=0.07). ‘Optimal’ rats also showed visible improvements in choice after acute CB1 antagonism (main dose x group interaction effect P=0.06), though both of

NU

these results are underpowered. However, outside of these subtle effects, acute modulation

MA

of the CB1 receptor and endocannabinoid system did not significantly alter risk-based decision making in the current study, an effect at odds with human literature in which acute

D

doses impaired decision making (Bolla et al., 2005). Additional data have shown acute THC

PT E

exposure in humans induced risky choice in laboratory tasks in frequent marijuana users (Lane and Cherek, 2002; Lane et al., 2005). However, these experiments were conducted in

CE

subjects with a history of marijuana use, which may have influenced the effect of THC on decision making parameters. Furthermore, other studies have shown that acute THC

AC

administration produced no impairments in recreational or heavy cannabis users (Ramaekers et al., 2006; Cousijn et al., 2013). The mixed effects seen in humans may be due to individual differences in activation of brain regions important for reward evaluation and decision making (Cousijn et al., 2013), suggesting differences in reward processing may explain discrepancies in responding to THC. Though our individual differences results are not significant, they would suggest this possibility merits further investigation with a greater cohort of rats that undergo both acute and chronic THC treatment. 21

ACCEPTED MANUSCRIPT Interestingly, previous work has shown acute THC administration can negatively impact other facets of cognition, including working memory (Morgan et al., 2010), cognitive flexibility as measured by task switching (Anderson et al., 2011), and emotional processing (Hindocha et al., 2015). Although decision making relies upon these cognitive faculties, it is possible the rGT did not actively recruit them at the time of the drug challenge. Clinical

PT

studies employing cognitive tasks require subjects to learn the paradigm within a session,

RI

during which time the individual is “exploring” rules and strategies (Daw et al., 2006). In

SC

contrast, we tested animals after acquisition of stable task performance, at which point knowledge of the contingencies is established and a choice strategy is being “exploited”

NU

(Daw et al., 2006; De Visser et al., 2011), which may be less vulnerable to acute cannabinoid

MA

manipulation. Furthermore, previous work investigating the effects of THC on motor impulsivity found that only repeated, chronic administration of the drug potentiated

D

behavioral disinhibition, whereas acute and limited chronic dosing had no effect (Irimia et

PT E

al., 2015). Therefore, acute CB1 activation and antagonism may not affect decision making preferences, but chronic manipulations could impair task performance. Future studies

CE

should take into account length of THC experience (i.e. acute versus chronic dosing), and time of administration (task acquisition versus performance) to further explore how CB1

AC

activity may influence risk-based decision making. Other task parameters, such as choice latencies, number of trials completed and omissions, were affected by rimonabant, THC and/or WIN55,212-2 in a similar fashion compared to previous data (Arguello and Jentsch, 2004; Pattij et al., 2007a; Silveira et al., 2017; Wiskerke et al., 2011b). Thus, suboptimal dose ranges seem unlikely to explain the lack of effects of these direct cannabinoid CB1 receptor agonists on choice behavior. Rather, these results suggest that acute modulation of CB1 receptor activity may affect distinct 22

ACCEPTED MANUSCRIPT aspects of cognition. Indeed, THC reduced impulsive responding in all animals, whereas the potent CB1 agonist WIN55,212-2 potentiated premature responding in risk-preferring rats, yet reduced it in optimal animals. Nonetheless, contrasting our previous data (Pattij et al., 2007a), rimonabant did not affect impulsive responding in the rGT. As discussed earlier the task structure of the current task differs from other tasks measuring impulsive responding,

PT

which may impact the interpretation of our current data in comparison to data obtained in

RI

the 5-choice serial reaction time task with cannabinoid ligands previously (Pattij et al.,

SC

2007a; Wiskerke et al., 2011b). Regardless, these results suggest greater CB1 activation may produce unique results dependent upon pre-existing differences in reward processing. It is

NU

well documented that CB1 receptor activation influences dopamine release (for review see

MA

Bloomfield et al., 2016), including within the nucleus accumbens, an area in which dopaminergic activity correlates with decision-making preferences (Sugam et al., 2012).

D

Furthermore, dopaminergic transmission within the nucleus accumbens has been shown to

PT E

influence impulsive responding (Cole and Robbins, 1989; Besson et al., 2010; Moreno et al., 2013; Pattij et al., 2007b). Although speculative, the high dose of THC given may reduce

CE

accumbal dopamine release in all animals (Bloomfield et al., 2016), whereas more direct activation of the CB1 receptor may differentially affect release depending on baseline

AC

differences in dopaminergic activity between risk-preferring and optimal animals. Future studies are required to determine the effect of CB1 receptor modulation on dopamine release in reward centers or interactions with other neurotransmitter systems such as glutamate (Colizzi et al., 2016), the latter which may also impact impulsive responding (Winstanley, 2011). Although WIN55,212-2 potentiated impulsive action in risk-preferring animals, it dose-dependently reduced motivation as exhibited by increased omissions, reduced number 23

ACCEPTED MANUSCRIPT of trials, and increased choice latency in all rats. The dissociation between measures of motivation and premature responding may reflect recruitment of separate networks affected by CB1 activation (Oleson et al., 2012). In conclusion, these data show acute modulation of the CB1 receptor has dissociated effects on various aspects of cognitive performance, potentially tapping into distinct

PT

mechanisms governing decision making, impulse control, and motivation. As such, acute

RI

challenges with cannabinoid ligands did not strongly affect decision making, yet impacted

SC

impulsive behavior and measures of motivation. The effects observed here add to an existing body of research exploring the role of the endocannabinoid system in cognition, and may

NU

shed light on how the recruitment of the CB1 receptor by exogenous cannabinoids may

MA

contribute to the expression of various behavioral phenotypes implicated in psychiatric

D

disease.

PT E

Acknowledgements

Jacqueline-Marie N. Ferland was supported by the Liberation Scholarship Program from the

Netherlands.

CE

Ministry of Foreign Affairs and the Ministry of Education, Culture and Science of the

AC

Angela M. Lee was supported by a Fulbright Program grant from the U.S. Department of State, which was funded through the Netherland-America Foundation.

Conflicts of interest There are no conflicts of interest.

24

ACCEPTED MANUSCRIPT References

Anderson, B.M., Rizzo, M., Block, R.I., Pearlson, G.D., O’Leary, D.S., 2011. Sex, drugs, and cognition: effects of marijuana. J. Psychoactive Drugs. 42, 413-424.

PT

Arguello, P.A., Jentsch, J.D., 2004. Cannabinoid CB1 receptor-mediated impairment of visuospatial

SC

RI

attention in the rat. Psychopharmacology (Berl). 177, 141-150.

Baarendse, P.J.J., Winstanley, C.A., Vanderschuren, L.J.M.J., 2013. Simultaneous blockade of

MA

task. Psychopharmacology (Berl). 225, 719-731.

NU

dopamine and noradrenaline reuptake promotes disadvantageous decision making in a rat gambling

Bari, A., Dalley, J.W., Robbins, T.W., 2008. The application of the 5-choice serial reaction time task for

PT E

D

the assessment of visual attentional processes and impulse control in rats. Nat. Protoc. 3, 759-767.

Bechara, A., 2005. Decision making, impulse control and loss of willpower to resist drugs: a

CE

neurocognitive perspective. Nat. Neurosci. 8, 1458-1463.

AC

Bechara, A., Damasio, A.R., Damasio, H., Anderson, S.W., 1994. Insensitivity to future consequences following damage to human prefrontal cortex. Cognition. 50, 7-15.

Besson, M., Belin, D., McNamara, R., Theobald, D.E., Castel, A., Beckett, V.L., Crittenden, B.M., Newman, A.H., Everitt, B.J., Robbins, T.W., Dalley, J.W., 2010. Dissociable control of impulsivity in rats by dopamine d2/3 receptors in the core and shell subregions of the nucleus accumbens. Neuropsychopharmacology. 35, 560-569.

25

ACCEPTED MANUSCRIPT Bizot, J.C., Cogrel, N., Massé, F., Chauvin, V., Brault, L., David, S., Trovero, F., 2015. D-amphetamine improves attention performance in adolescent Wistar, but not in SHR rats, in a two-choice visual discrimination task. Psychopharmacology (Berl). 232, 3269-3286.

Bloomfield, M.A.P., Ashok, A.H., Volkow, N.D., Howes, O.D., 2016. The effects of Δ9-

PT

tetrahydrocannabinol on the dopamine system. Nature. 539, 369-377.

SC

NU

in abstinent marijuana users. NeuroImage. 26, 480-492.

RI

Bolla, K.I., Eldreth, D.A., Matochik, J.A., Cadet, J.L., 2005. Neural substrates of faulty decision-making

Cella, M., Dymond, S., Cooper, A., 2010. Impaired flexible decision-making in Major Depressive

MA

Disorder. J. Affect. Disord. 124, 207-210.

D

Cocker, P.J., Dinelle, K., Kornelson, R., Sossi, V., Winstanley, C.A., 2012. Irrational choice under

PT E

uncertainty correlates with lower striatal D(2/3) receptor binding in rats. J. Neurosci. 32, 1545015457.

CE

Cole, B.J., Robbins, T.W., 1989. Effects of 6-hydroxydopamine lesions of the nucleus accumbens septi

AC

on performance of a 5-choice serial reaction time task in rats: implications for theories of selective attention and arousal. Behav. Brain Res. 33, 165-179.

Colizzi, M., McGuire, P., Pertwee, R.G., Bhattacharyya, S., 2016. Effect of cannabis on glutamate signalling in the brain: A systematic review of human and animal evidence. Neurosci. Biobehav. Rev. 64, 359-381.

26

ACCEPTED MANUSCRIPT Cousijn, J., Wiers, R.W., Ridderinkhof, K.R., van den Brink, W., Veltman, D.J., Porrino, L.J., Goudriaan, A.E., 2013. Individual differences in decision making and reward processing predict changes in cannabis use: a prospective functional magnetic resonance imaging study. Addict. Biol. 18, 10131023.

PT

Daw, N.D., O’Doherty, J.P., Dayan, P., Seymour, B., Dolan, R.J., 2006. Cortical substrates for

RI

exploratory decisions in humans. Nature. 441, 876-879.

SC

De Visser, L., Homberg, J.R., Mitsogiannis, M., Zeeb, F.D., Rivalan, M., Fitoussi, A., Galhardo, V., Van

NU

den Bos, R., Winstanley, C.A., Dellu-Hagedorn, F., 2011. Rodent versions of the iowa gambling task:

MA

opportunities and challenges for the understanding of decision-making. Front. Neurosci. 5, 109.

De Vries, T.J., Shaham, Y., Homberg, J., Crombag, H., Schuurman, K., Dieben, J., Vanderschuren,

D

L.J.M.J., Schoffelmeer, A.N.M., 2001. A cannabinoid mechanism in relapse to cocaine seeking. Nat.

PT E

Med. 7, 1151-1154.

De Vries, T.J., De Vries, W., Janssen, M.C., Schoffelmeer, A.N., 2005. Suppression of conditioned

AC

161, 164-168.

CE

nicotine and sucrose seeking by the cannabinoid-1 receptor antagonist SR141716A. Behav. Brain Res.

Di Ciano, P., Pushparaj, A., Kim, A., Hatch, J., Masood, T., Ramzi, A., Khaled, M.A., Boileau, I., Winstanley, C.A., Le Foll, B., 2015. The impact of selective dopamine D2, D3 and D4 ligands on the rat gambling task. PLoS One. 10, e0136267.

Ferland, J.N., Winstanley, C.A., 2017. Risk-preferring rats make worse decisions and show increased incubation of craving after cocaine self-administration. Addict. Biol. 22, 991-1001.

27

ACCEPTED MANUSCRIPT

French, E.D., 1997. Delta9-Tetrahydrocannabinol excites rat VTA dopamine neurons through activation of cannabinoid CB1 but not opioid receptors. Neurosci. Lett. 226, 159-162.

Gueye, A.B., Trigo, J.M., Vemuri, K.V., Makriyannis, A., Le Foll, B., 2016. Effects of various cannabinoid

PT

ligands on choice behaviour in a rat model of gambling. Behav. Pharmacol. 27, 258-269.

RI

Haj-Dahmane, S., Shen, R.Y., 2009. Endocannabinoids suppress excitatory synaptic transmission to

SC

dorsal raphe serotonin neurons through the activation of presynaptic CB1 receptors. J. Pharmacol.

NU

Exp. Ther. 331, 186-196.

MA

Hindocha, C., Freeman, T.P., Schafer, G., Gardener, C., Das, R.K., Morgan, C.J. et al., 2015. Acute effects of delta-9-tetrahydrocannabinol, cannabidiol, and their combination on facial emotion

D

recognition: a randomised, double-blind, placebo-controlled study in cannabis users. Eur.

PT E

Neuropsychopharmacol. 25, 325-334.

Irimia, C., Polis, I.Y., Stouffer, D., Parsons, L.H., 2015. Persistent effects of chronic Δ9-THC exposure

AC

CE

on motor impulsivity in rats. Psychopharmacology (Berl). 232, 3033-3043.

Lane, S.D., Cherek, D.R., 2002. Marijuana effects on sensitivity to reinforcement in humans. Neuropsychopharmacology. 26, 520-529.

Lane, S.D., Cherek, D.R., Tcheremissine, O.V., Lieving, L.M., Pietras, C.J., 2005. Acute marijuana effects on human risk taking. Neuropsychopharmacology. 30, 800–809.

28

ACCEPTED MANUSCRIPT Lawn, W., Freeman, T.P., Pope, R.A., Joye, A., Harvey, L., Hindocha, C., Mokrysz, C., Moss, A., Wall, M.B., Bloomfield, M.A., Das, R.K., Morgan, C.J., Nutt, D.J., Curran, H.V. 2016. Acute and chronic effects of cannabinoids on effort-related decision-making and reward learning: an evaluation of the cannabis ‘amotivational’ hypothesis. Psychopharmacology (Berl)., 233 (19-20): 3537-52.

PT

McDonald, J., 2009. Handbook of biological statistics, 2nd ed. Sparky House Publishing, Baltimore,

RI

MD, USA.

SC

Mitchell, M.R., Weiss, V.G., Beas, B.S., Morgan, D., Bizon, J.L., Setlow, B., 2014. Adolescent risk taking,

NU

cocaine self-administration, and striatal dopamine signaling. Neuropsychopharmacology. 39, 955-

MA

962.

Moreno, M., Economidou, D., Mar, A.C., Lopez-Granero, C., Caprioli, D., Theobald, D.E., Fernando, A.,

D

Newman, A.H., Robbins, T.W., Dalley, J.W., 2013. Divergent effects of D2/3 receptor activation in the

PT E

nucleus accumbens core and shell on impulsivity and locomotor activity in high and low impulsive rats. Psychopharmacology (Berl), 228: 19-30.

CE

Morgan, C.J., Schafer, G., Freeman, T.P., Curran, H.V., 2010. Impact of cannabidiol on the acute

285-290.

AC

memory and psychotomimetic effects of smoked cannabis: naturalistic study. Br. J. Psychiatry. 197,

Oleson, E.B., Beckert, M.V., Morra, J.T., Lansink, C.S., Cachope, R., Abdullah, R.A., Loriaux, A.L., Schetters, D., Pattij, T., Roitman, M.F., Lichtman, A.H., Cheer, J.F., 2012. Endocannabinoids shape accumbal encoding of cue-motivated behavior via CB1 receptor activation in the ventral tegmentum. Neuron. 73, 360-373.

29

ACCEPTED MANUSCRIPT Parsons, L.H., Hurd, Y.L., 2015. Endocannabinoid signalling in reward and addiction. Nat. Rev. Neurosci. 16, 579-594.

Pattij, T., Janssen, M.C., Schepers, I., Gonzalez-Cuevas, G., De Vries, T.J., Schoffelmeer, A.N., 2007a. Effects of the cannabinoid CB(1) receptor antagonist rimonabant on distinct measures of impulsive

PT

behavior in rats. Psychopharmacology (Berl). 193, 85-96.

RI

Pattij, T., Janssen, M.C., Vanderschuren, L.J., Schoffelmeer, A.N., Van Gaalen, M.M., 2007b.

SC

Involvement of dopamine D1 and D2 receptors in the nucleus accumbens core and shell in inhibitory

NU

response control. Psychopharmacology (Berl). 191, 587-598.

MA

Pedersen, A., Goder, R., Tomczyk, S., Ohrmann, P. 2017. Risky decision-making under risk in

D

schizophrenia: A deliberate choice? J Behav Ther Exp Psychiatry, 56: 57-64.

PT E

Piomelli, D., 2003. The molecular logic of endocannabinoid signalling. Nat Rev Neurosci. 4, 873-884.

Potenza, M.N., 2009. The importance of animal models of decision making, gambling, and related

AC

2624.

CE

behaviors: implications for translational research in addiction. Neuropsychopharmacology. 34, 2623-

Ramaekers, J.G., Moeller, M.R., Van Ruitenbeek, P., Theunissen, E.L., Schneider, E., Kauert, G., 2006. Cognition and motor control as a function of Delta9-THC concentration in serum and oral fluid: limits of impairment. Drug Alc. Depend. 85, 114-122.

Renard, J., Krebs, M.O., Le Pen, G., Jay, T.M., 2014. Long-term consequences of adolescent cannabinoid exposure in adult psychopathology. Front. Neurosci 8, 361.

30

ACCEPTED MANUSCRIPT

Sanchis-Segura, C., Cline, B.H., Marsicano, G., Lutz, B., Spanagel, R., 2004. Reduced sensitivity to reward in CB1 knockout mice. Psychopharmacology (Berl). 176, 223-232.

Silveira, M.M., Malcolm, E., Shoaib, M., Winstanley, C.A., 2015. Scopolamine and amphetamine

PT

produce similar decision-making deficits on a rat gambling task via independent pathways. Behav.

M.M.,

Adams,

W.K.,

Morena,

M.,

Hill,

M.N.,

Winstanley,

SC

Silveira,

RI

Brain Res. 281, 86-95.

C.A.,

2017.

Δ9-

NU

Tetrahydrocannabinol decreases willingness to exert cognitive effort in male rats. J. Psychiatry

MA

Neurosci. 42, 131-138.

Sugam, J.A., Day, J.J., Wightman, R.M., Carelli, R.M., 2012. Phasic nucleus accumbens dopamine

PT E

D

encodes risk-based decision-making behavior. Biol. Psychiatry. 71, 199-205.

United Nations, 2015. United Nations office on drugs and crime world drug report. United Nations,

CE

New York, USA.

AC

Van Gaalen, M.M., Brueggeman, R.J., Bronius, P.F., Schoffelmeer, A.N., Vanderschuren, L.J., 2006. Behavioral disinhibition requires dopamine receptor activation. Psychopharmacology (Berl). 187, 7385.

Van Hell, H.H., Jager, G., Bossong, M.G., Brouwer, A., Jansma, J.M., Zuurman, L., van Gerven, J., Kahn, R.S., Ramsey, N.K. 2012. Involvement of the endocannabinoid system in reward processing in the human brain. Psychopharmacology (Berl)., 219(4): 981-990.

31

ACCEPTED MANUSCRIPT Winstanley, C.A., 2011. The utility of rat models of impulsivity in developing pharmacotherapies for impulse control disorders. Br. J. Pharmacol. 164, 1301-1321.

Winstanley, C.A., Clark, L., 2016. Translational models of gambling-related decision-making. Cur

PT

Topics Behav. Neurosci. 28, 93-120.

Wiskerke, J., Schetters, D., Van Es, I.E., Van Mourik, Y., Den Hollander, B.R., Schoffelmeer, A.N., Pattij,

RI

T., 2011a. μ-Opioid receptors in the nucleus accumbens shell region mediate the effects of

NU

SC

amphetamine on inhibitory control but not impulsive choice. J. Neurosci. 31, 262-272.

Wiskerke J, Stoop N, Schetters D, Schoffelmeer AN, Pattij T (2011b) Cannabinoid CB1 receptor

MA

activation mediates the opposing effects of amphetamine on impulsive action and impulsive choice.

D

PLoS One 6:e25856.

PT E

Wiskerke, J., Van Mourik, Y., Schetters, D., Schoffelmeer, A.N., Pattij, T., 2012. On the role of cannabinoid CB1- and μ-opioid receptors in motor impulsivity. Front Pharmacol. 3, 108.

CE

Zeeb, F.D., Robbins, T.W., Winstanley, C.A., 2009. Serotonergic and dopaminergic modulation of

2343.

AC

gambling behavior as assessed using a novel rat gambling task. Neuropsychopharmacology 34, 2329-

Zeeb, F.D., Winstanley, C.A., 2011. Lesions of the basolateral amygdala and orbitofrontal cortex differentially affect acquisition and performance of a rodent gambling task. The Journal of neuroscience : the official journal of the Society for Neuroscience 31, 2197-2204.

32

ACCEPTED MANUSCRIPT Zeeb, F.D., Winstanley, C.A., 2013. Functional disconnection of the orbitofrontal cortex and basolateral amygdala impairs acquisition of a rat gambling task and disrupts animals' ability to alter decision-making behavior after reinforcer devaluation. The Journal of neuroscience : the official journal of the Society for Neuroscience 33, 6434-6443.

PT

Zhang, L., Wang, K., Zhu, C., Yu, F., Chen, X., 2015. Trait anxiety has effect on decision making under

AC

CE

PT E

D

MA

NU

SC

RI

ambiguity but not decision making under risk. PloS one 10, e0127189.

33

ACCEPTED MANUSCRIPT Table 1. rGT choice parameters, reward and punishment schedules, and nose-poke hole locations. Choice

Reward (# pellets)

Punishment (s)

P (punishment)

Nosepoke hole version A

version B

1

5

0.1

1

2

P2

2

10

0.2

4

5

P3

3

30

0.5

P4

4

40

0.6

PT

P1

4

2

1

NU

SC

RI

5

Table 2. Auxiliary parameters of baseline performance in the rGT averaged over three

Parameter

MA

baseline sessions from both cohorts. ‘Optimal’ rats

0.2 (0.1)

Premature responses

Pellets earned Choice latency (s)

AC

Collection latency (s)

0.3 (0.2)

9.6 (2.1)

8.4 (2.9)

98.9 (5.4)

66.9 (4.0)**

145.8 (8.2)

97.3 (6.6)**

1.2 (0.09)

1.2 (0.2)

1.3 (0.5)

0.6 (0.3)

CE

Trials

PT E

Omissions

n=12

D

n=20

‘Risk-preferring’ rats

Values expressed as average (SEM). **P<0.005 compared to ‘optimal’ rats

34

ACCEPTED MANUSCRIPT Table 3. Auxiliary parameters after rimonabant administration.

Choice Preference

Parameter

Rimonabant Dose

3 mg/kg

Omissions

0.1 (0.1)

0 (0.0)

0.2 (0.1)

0.3 (0.2)

Trials

108.3 (7.5)

107.4 (9.0)

111.0 (7.9)

97.6 (8.1)

Choice latency (s)

0.9 (0.1)

1.0 (0.1)

0.9 (0.1)

1.1 (0.1)*

Collection latency (s)

0.7 (0.04)

1.3 (0.6)

0.8 (0.1)

0.8 (0.1)

Omissions

0.3 (0.3)

0 (0.0)

0.3 (0.2)

0.3 (0.2)

Trials

67.5 (6.3)

73.7 (7.9)

64.0 (3.4)

61.8 (4.6)

Choice latency (s)

0.9 (0.2)

1.1 (0.2)

1.1 (0.3)

1.2 (0.2)*

Collection latency (s)

0.5 (0.1)

0.5 (0.1)

0.6 (0.2)

1.4 (0.9)

NU

SC

RI

1 mg/kg

MA

Risk-preferring

0.3 mg/kg

D

Optimal

Vehicle

PT

Group

AC

CE

PT E

Values expressed as averages per dose (SEM). *P<0.05 compared to vehicle.

35

ACCEPTED MANUSCRIPT

Table 4. Auxiliary parameters after WIN 55,212-2 administration. Choice Preference

Parameter

WIN 55,212-2 Dose

Group 3 mg/kg

Omissions

0 (0.0)

0.1 (0.1)

2.9 (1.2)*

5.1 (2.2)*

Trials

103.9 (7.8)

104.8 (7.6)

Choice latency (s)

0.9 (0.1)

0.8 (0.1)

Collection latency (s)

0.7 (0.04)

0.8 (0.1)

Omissions

0 (0.0)

Trials

65.8 (3.9)

Choice latency (s) Collection latency (s)

PT

1 mg/kg

72.4 (8.2)*

1.6 (0.2)

2.0 (0.2)*

1.0 (0.1)

1.0 (0.1)

0 (0.0)

0 (0.0)

5.5 (4.9)*

60.7 (2.0)

60.3 (3.7)*

35.8 (10.7)*

1.0 (0.2)

1.0 (0.2)

0.9 (0.1)

1.4 (0.4)*

0.5 (0.02)

0.5 (0.02)

0.5 (0.03)

0.6 (0.1)

NU

SC

RI

74.6 (12.5)*

D

Risk-preferring

0.3 mg/kg

MA

Optimal

Vehicle

AC

CE

PT E

Values expressed as averages per dose (SEM). *P<0.05 compared to vehicle.

36

ACCEPTED MANUSCRIPT Table 5. Auxiliary parameters after THC administration. Choice Preference

Parameter

THC Dose

2 mg/kg

Omissions

0 (0.0)

0.1 (0.1)

0.2 (0.2)

2.3 (1.5)

Trials

103.8 (8.4)

108.0 (8.6)

99.1 (10.3)

83.9 (8.5)

Choice latency (s)

0.9 (0.1)

0.9 (0.1)

1.1 (0.1)

1.7 (0.2)*

Collection latency (s)

0.7 (0.04)

0.8 (0.04)

0.9 (0.1)

0.9 (0.1)

Omissions

0.3 (0.3)

0.2 (0.2)

0 (0.00)

0.2 (0.2)

Trials

69.2 (3.6)

63.7 (4.3)

66.5 (3.9)

71.0 (7.0)

Choice latency (s)

1.1 (0.3)

1.1 (0.2)

1.0 (0.2)

1.0 (0.1)

Collection latency (s)

0.5 (0.1)

0.5 (0.03)

0.5 (0.04)

0.6 (0.1)

SC

RI

1 mg/kg

NU

Risk-preferring

0.5 mg/kg

MA

Optimal

Vehicle

PT

Group

AC

CE

PT E

D

Values expressed averages per dose (SEM). *P<0.05 compared to vehicle.

37

ACCEPTED MANUSCRIPT (Table 6. Auxiliary parameters after FAAH synthesis inhibitor URB597 administration. Parameter

URB597 Dose Vehicle

0.3 mg/kg

0.5 (0.3)

0.4 (0.2)

92.4 (15.2)

87.2 (14.6)

Choice latency (s)

1.3 (0.2)

1.3 (0.2)

Collection latency (s)

0.9 (0.2)

0.9 (0.2)

Omissions

RI

Choice Preference Group

0.2 (0.1)

0.2 (0.1)

Omissions

Risk-preferring

PT

Trials

SC

Optimal

Trials

Collection latency (s)

61.5 (13.8)

1.2 (0.3)

1.2 (0.3)

0.6 (0.1)

0.6 (0.1)

AC

CE

PT E

D

MA

Values expressed as averages per dose SEM).

NU

Choice latency (s)

62.8 (13.8)

38

ACCEPTED MANUSCRIPT 39 Table 7. Auxiliary parameters after amphetamine administration. Choice

Parameter

Amphetamine Dose

Preference

0 (0.0)

0.9 (0.6)

82.4 (13.4)

56.2 (10.5)

Choice latency (s)

1.2 (0.2)

1.4 (0.3)

Collection latency (s)

0.9 (0.1)

Omissions

4.8 (3.2)

0.2 (0.1)

58.3 (12.3)

49.9 (10.1)

1.2 (0.3)

1.0 (0.2)

0.6 (0.1)

0.6 (0.2)

Omissions

Riskpreferring

1.3 (0.5)**

NU

Trials

PT

1 mg/kg

RI

Optimal

Vehicle

SC

Group

MA

Trials Choice latency (s)

D

Collection latency (s)

AC

CE

PT E

Values expressed as averages per dose (SEM). *P<0.05 and **P<0.005 compared to respective vehicle.

ACCEPTED MANUSCRIPT 40

Legends to the figures

Figure 1 Average choice preference over three baseline sessions (A) for the different pellet options in the rGT in optimal rats (N = 20; white bars) and risk-preferring rats (N = 12; black bars); individual differences

PT

in average choice score over three baseline sessions (B) and choice score across three different

SC

RI

baseline sessions (C).

Figure 2

NU

Acute challenges with the cannabinoid CB1 receptor antagonist rimonabant neither shift choice preference (A,B), nor choice score (C) nor premature responding (D) in ‘optimal’ and ‘risk-preferring’

MA

rats.

D

Figure 3

PT E

Acute challenges with the cannabinoid CB1 receptor agonist THC shift neither choice preference (A,B) nor choice score (C), but decrease premature responding (D) in ‘optimal’ and ‘risk-preferring’ rats.

CE

Challenges with the cannabinoid CB1 receptor agonist WIN55,212-2 does not shift choice preference (E,F) or choice score (G), but exert opposite effects on premature responding (H) in ‘optimal’ and

AC

‘risk-preferring’ rats.

*P<0.05 compared to vehicle. #P<0.05 and ##P<0.005 ‘optimal’ rats versus ‘risk-preferring’ rats.

Figure 4 Acute challenges with the FAAH inhibitor URB597 (0.3 mg/kg) neither shift choice preference (A), nor choice score (B) nor premature responding (C) in ‘optimal’ and ‘risk-preferring’ rats.

ACCEPTED MANUSCRIPT 41 Figure 5 Acute challenges with the psychostimulant amphetamine (1.0 mg/kg) shifts choice preferences in both ‘optimal’ rats and ‘risk-preferring’ rats (A), without affecting choice score (B) or premature

AC

CE

PT E

D

MA

NU

SC

RI

PT

responding (C). *P<0.05, vehicle versus amphetamine.

ACCEPTED MANUSCRIPT 42

AC

CE

PT E

D

MA

NU

SC

RI

PT

Figure 1

ACCEPTED MANUSCRIPT 43

AC

CE

PT E

D

MA

NU

SC

RI

PT

Figure 2

ACCEPTED MANUSCRIPT 44

AC

CE

PT E

D

MA

NU

SC

RI

PT

Figure 3

ACCEPTED MANUSCRIPT 45

AC

CE

PT E

D

MA

NU

SC

RI

PT

Figure 4

ACCEPTED MANUSCRIPT 46

AC

CE

PT E

D

MA

NU

SC

RI

PT

Figure 5

ACCEPTED MANUSCRIPT 47

Highlights: stable individual differences in choice preference in the rat gambling task were detected in male Wistar rats

-

acute challenges with the psychostimulant drug amphetamine promote disadvantageous choice

-

acute cannabinoid CB1 receptor modulation has dissociable effects on decision making, motor impulsivity and motivation

-

inhibition of fatty acid amide hydrolase, an enzyme for degradation of the endogenous cannabinoid anandamide, does not affect cognitive performance in the rat gambling task

AC

CE

PT E

D

MA

NU

SC

RI

PT

-