Serotonergic and dopaminergic control of impulsivity in gambling disorder

Abstract Gambling disorder (GD) is major public health issue. The disorder is often characterized by elevated impulsivity with evidence from analogous substance use disorders underlining prominent roles of brain monoamines in addiction susceptibility and outcome. Critically, GD allows the study of addiction mechanisms without the confounder of the effects of chronic substances. Here, we assessed the roles of striatal dopamine transporter binding and extrastriatal serotonin transporter binding in GD as a function of impulsivity using [123I]FP‐CIT SPECT imaging in 20 older adults with GD (DSM‐5 criteria; mean age 64 years) and 40 non‐GD age‐ and sex‐matched controls. We focused on GD in older individuals because there are prominent age‐related changes in neurotransmitter function and because there are no reported neuroimaging studies of GD in older adults. Volume‐of‐interest‐based and voxelwise analyses were performed. GD patients scored clearly higher on impulsivity and had higher tracer binding in the ventromedial prefrontal cortex than controls (p < 0.001), likely reflecting serotonin transporter activity. The binding in the medial prefrontal cortex positively correlated with impulsivity over the whole sample (r = 0.62, p < 0.001) as well as separately in GD patients (r = 0.46, p = 0.04) and controls (r = 0.52, p < 0.001). Striatal tracer binding, reflecting dopamine transporter activity was also positively correlated with impulsivity but showed no group differences. These findings highlight the role of prefrontal serotonergic function in GD and impulsivity. They identify cerebral coordinates of a potential target for neuromodulation for both GD and high impulsivity, a core phenotypic dimensional cognitive marker in addictions.

(BCLC), a Canadian Crown Corporation. The Province of BC government and the BCLC had no role in the preparation of this article and impose no constraints on publishing. LC has received a speaker/travel honorarium from the National Association for Gambling Studies (Australia) and the National Center for Responsible Gaming (US) and has received fees for academic services from the National Center for Responsible Gaming (US), GambleAware (UK) and Gambling Research Exchange Ontario (Canada). He has not received any further direct or indirect payments from the gambling industry or groups substantially funded by gambling. He has received royalties from Cambridge Cognition Ltd. relating to neurocognitive testing. Dr. Voon  well as separately in GD patients (r = 0.46, p = 0.04) and controls (r = 0.52, p < 0.001). Striatal tracer binding, reflecting dopamine transporter activity was also positively correlated with impulsivity but showed no group differences. These findings highlight the role of prefrontal serotonergic function in GD and impulsivity. They identify cerebral coordinates of a potential target for neuromodulation for both GD and high impulsivity, a core phenotypic dimensional cognitive marker in addictions.

K E Y W O R D S
gambling disorder, impulsivity, SPECT

| INTRODUCTION
Gambling disorder (GD), characterized by persistent, recurrent maladaptive patterns of gambling behaviour, is a public health concern worldwide. 1 The increasing recognition of GD is combined with the limited success of current treatment approaches, 2,3 and there are no pharmacological agents approved in any country for the treatment of GD. 4 A key task in the field is therefore the translation of the underlying neurobiological mechanisms into better prevention and treatment strategies for GD. 4,5 The neurobiology of GD remains largely unclear but brain circuits involving ventral prefrontal, ventral striatal and limbic brain regions seem to contribute to reward-related decision-making biases that drive uncontrolled gambling. 5 Multiple neurotransmitter systems in these circuits have been implicated, but the brain monoamines dopamine (3,4-dihydroxyphenethylamine, DA) and serotonin (5-hydroxytryptamine, 5-HT) play major roles in impulsive and compulsive behaviours. Studies of DA function in reward-related processes have been instrumental in our understanding of compulsive drug-seeking behaviour, 6 but further adaptations in the 5-HT system are also implicated in susceptibility to compulsive drug use. 7 In healthy individuals, and with relevance to GD, the 5-HT system seems to modulate decision-making under risk. 8 However, DA or 5-HT neuroimaging studies in GD are scarce and have shown inconsistent results. For DA function, studies have reported no changes in DA D 2/3 receptor binding in GD, [9][10][11][12][13] increased striatal DA synthesis capacity, 14 no change in DA synthesis capacity, 15 and decreased striatal DA transporter (DAT) binding. 16 The lack of data is even more notable for 5-HT neuroimaging with only two reported studies in GD, one finding no evidence of alterations in 5-HT 1B receptor binding in GD 17 and the other reporting the same for 5-HT transporter (SERT) binding. 18 Moreover, samples in GD neuroimaging studies have generally been small (total number of GD Subjects n = 12-18), 19 which could explain the inconsistencies and negative results.
Although GD is highly heterogeneous in terms of phenotypic characteristics, patients are characterized by high impulsivity across various subdomains of impulsivity, 20 and impulsivity is also associated with psychiatric comorbidity (e.g., ADHD). In substance use disorders, increased impulsivity is a robust phenomenon across different addictions, is associated with clinical outcomes, and can be a pre-existing risk for vulnerability and dependence. 21 From a therapeutic point of view, there is increasing evidence that supports pharmacological and behavioural treatments for impulsivity in substance use disorders, 22 but the underlying mechanisms and treatment targets are poorly defined, particularly for high impulsiveness in GD.
In the present study, we aimed to fill the knowledge gap concerning monoaminergic function in GD with a focus on impulsivity, a prominent abnormality in GD patients. For the neuroimaging tracer, we used [ 123 I]FP-CIT, a cocaine analogue that has high affinity for DAT in the striatum but also a moderate affinity for SERT outside the striatum, thus enabling a simultaneous investigation of two monoamine transporters in one imaging session. [23][24][25] We focused on GD in older individuals because there are prominent age-related changes in neurotransmitter function that may be relevant to pharmacological treatments and addictions 26 and because there are no reported neuroimaging studies of GD in older adults. Our sample size represents the largest GD serotonergic or dopaminergic neuroimaging study to date.

| Subjects
Twenty individuals with GD and 40 non-GD controls were recruited (Table 1)  Sixteen out of the 20 subjects with GD (80%) expressed a preference for slot machine gambling (remaining four subjects: horse race betting, bingo, roulette or missing data). None of the GD or control subjects were using medications affecting the dopaminergic system.
One individual with GD and one control received 5-to 10-mg/day escitalopram for a mild mood disorder without a diagnosis of major depression; no other selective serotonin reuptake inhibitors (SSRIs) were used by the participants. None had prior or current clinically relevant neurological conditions, and there were no other prior or current psychiatric diagnoses.

| Demographic and clinical characteristics
In addition to group-differences in gambling-and impulsivity-related variables, differences between the GD and control groups were observed in BMI (higher in GD), duration of formal education (shorter in GD) and BDI depression scores (higher in GD) ( Table 1).

| Imaging
In the VOI analysis, no differences were observed between GD and controls in caudate nucleus or putamen tracer binding (Table 1 and Figure 1). In the voxelwise analysis, a group difference was observed in the ventromedial prefrontal cortex (vmPFC), with peak voxels at À24 42 26, À26 38 16 and 12 42 À4 (Figure 1). Tracer binding in the vmPFC was 10.7% higher in GD patients than in controls (p < 0.001,

| Impulsivity
The BIS-11 total score correlated positively with tracer binding in the vmPFC cluster region in all subjects (GD and controls combined, r = 0.62, p < 0.001) and separately in healthy controls (r = 0.52, p < 0.001) and GD patients (r = 0.46, p = 0.043) (Figure 2). There was no significant difference in the correlation coefficients between healthy controls and GD patients (p = 0.79).

| Clinical characteristics
Striatal or vmPFC tracer binding did not correlate with SOGS scores The second main finding of this study was the positive relationship between vmPFC tracer binding and impulsivity, over the whole sample and separately in GD and controls. In the whole sample and in controls, the statistical significance was high (p < 0.001). Impulsivity, referring to premature, unduly risky, poorly conceived actions, is strongly associated with GD as well as substance use disorders. 20,22,38,39 Data from a number of earlier studies suggest an important role of the prefrontal cortex in controlling multiple types of impulsivity 40 with the vmPFC, particularly implicating delay discounting, risk taking and goal-directed control. 41 Rodent models have identified enhanced SERT function in the orbitofrontal cortex and medial prefrontal regions as a potential mechanism that underlies impulsive behaviour contributing to reward processing. 42,43 The literature linking serotonergic transmission to 'waiting' impulsivity or premature responding is extensive but mostly based on animal studies.
'Waiting' impulsivity in humans implicates the vmPFC or human analogue of the prelimbic cortex 44 and tryptophan depletion lowering central serotonergic levels in healthy humans also results in greater 'waiting' impulsivity. 45 Genetic or pharmacological disruption of SERT or administration of agents that elevate synaptic 5-HT levels reduce premature responses in animals, whereas depletion of brain 5-HT has the opposite effect. 46 In humans, tryptophan depletion studies have indicated an inverse relationship between brain serotonin levels and shown some promise in modulating striatal DAT function in GD. 51 Third, we found no evidence for alterations in DAT binding in GD compared to controls, although striatal DAT binding was also positively correlated with impulsivity scores in both groups, which is in line with a previous [ 123 I]FP-CIT SPECT study in healthy subjects, 52 although an opposite correlation 53 and negative results 54 have also been reported. One earlier study reported an 11%-13% striatal DAT binding decrease in GD compared to controls. 16 The reasons for differences in the results between the earlier smaller study (GD n = 15 vs. controls n = 17) 16 and the present study are not clear. However, in the earlier study, all but one subject were males, subjects were considerably younger and impulsivity scores were somewhat higher. The dopamine system shows substantial age-related changes (e.g., in D2/3 binding), but older age is increasingly recognized as a vulnerable window for disordered gambling. 55 In addition, some work has indicated nonlinear (i.e., U-shaped) relationships between dopamine D2/3 receptor binding and impulsivity, both in GD and healthy participants, such that both high and low extremes of impulsivity may be associated with reduced dopamine transmission. 9,56 It is possible that these factors explain the differences compared to the present study, although we did not observe reduced DAT binding in our male GD patients compared to male controls (data not shown) and the group difference in striatal binding values was clearly nonsignificant in our study without any trend-level findings. The mean age of our sample underlines the issue of lack of striatal differences and higher prefrontal binding. First, the results are unlikely to be related to age-/GDassociated brain atrophy, since atrophy would have induced partial volume effects and would have led to lower binding ratios in GD. Second, there are prior reports that have linked substance use disorders with acceleration of normal ageing, possibly driven by immunosenescence and accelerated telomere shortening. [57][58][59] If this theory applies to GD, one would expect that the degeneration of DAT and SERT systems would be particularly prominent in aged individuals with long-term GD. Instead, transporter function was normal (DAT), arguing against accelerated cellular senescence in GD. It is also important to note that although we did not observe an association between the dysfunction in neurotransmission and the duration of GD behaviour, the transition time between problematic gaming and pathological gambling may be longer in older individuals with GD. 60  medications or Parkinsonian pathology. 61 Finally, genetic susceptibility may also be important, as has been hypothesized in aripiprazoleinduced GD via genetic polymorphism on the dopamine D2 receptor. 62 To summarize, we have demonstrated increased prefrontal SERT function in GD associated with individual differences in impulsivity.
Although we focused on GD, the findings may have a bearing on the understanding of neurotransmitter function in other behavioural and substance addictions and particularly their relationship with impulsive behaviour. An implication of this is the possibility that pathologically increased impulsive behaviour may be modulated in humans by interventions that target 5-HT function in the vmPFC.
Thus, although the results should be replicated independently with a larger number of younger GD patients, our findings highlight potential novel targets for neuromodulation both for GD and the dimension of impulsivity.

ACKNOWLEDGEMENTS
The staff of the Department of Nuclear Medicine, Turku University Hospital, is gratefully acknowledged.

CONFLICT OF INTEREST
There are no relevant conflicts of interest.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.