Candidate Gene Studies in Problem Gambling Dr. James L

Candidate Gene Studies in Problem Gambling

Dr. James L. Kennedy, MD, FRCP(C) Neuroscience Research Department Centre for Addiction and Mental Health

Dr. Nigel Turner, PhD Centre for Addiction and Mental Health

Dr. Daniela S. S. Lobo, MD, PhD Neuroscience Research Department Centre for Addiction and Mental Health (Collaborator in this study)

April 1, 2007

Final report prepared for the Ontario Problem Gambling Research Centre (OPGRC)

Disclaimer: Opinions expressed in this final report are those of the investigator(s), and do not necessarily represent the views of the OPGRC. 2

Table of Contents

Acknowledgements …………………………………………………………… 3

Abstract …………………………………………………………… 4

Introduction …………………………………………………………… 5

Research Hypotheses …………………………………………………………… 13

Method …………………………………………………………… 13

Results …………………………………………………………… 15

Discussion …………………………………………………………… 26

Study Limitations …………………………………………………………… 28

References …………………………………………………………… 31

Acknowledgements

This report was funded by a grant from the Ontario Problem Gambling Research Centre (OPGRC). The ideas expressed are those of the authors and do not necessarily reflect those of either the OPGRC or the Centre for Addiction and Mental Health.

Abstract

A growing body of evidence is demonstrating the susceptibility to become a pathological gambler has a relatively large genetic component. When exposed to the same environmental gambling opportunities, only some individuals develop symptoms of problem or pathological gambling (PG), which supports the presence of a biological susceptibility to develop the disorder in addition to environmental factors. Studies investigating the influence of genetic factors on the development of PG have provided strong evidence towards the heritability of PG vulnerability. Based on the current literature the monoaminergic systems and, in particular, the dopaminergic pathway, continue to represent the principal biological component involved in the etiology of PG. Previous molecular genetic studies in PG have been conducted to test dopamine and serotonin system genes as candidates, and concluded that variation in these genes may determine an increased risk of suffering from PG, but interpretation of these results remain limited by the sample sizes used and by the technology available at the time these studies were performed. The current study is the third part of our continuing effort to understand the genetic causes of PG. The first part, Special Problems in gambling: Attention Deficit Hyperactivity Disorder (ADHD) and pathways to problem gambling, focused on assessing the relative importance of emotional, experiential/behavioral, and physiological (genetic) aspects of the disorder. The study recruited and characterized PGs, as well as a control sample of non-problem, but active, social gamblers. The second part, Identification of genetic risk factors for pathological gambling (Kennedy, Muglia, Jain, & Turner, 2004), collected data from a control sample of infrequent or non-gamblers to act as a comparison to problem and non-problem social gamblers. Analysis of these samples showed indicators of association with genetic variants in the serotonin transporter, dopamine D1 and D4 receptors, and monoamine oxidase A. A larger sample was required to dissect and understand these findings. The aim of the current project was to collect a larger sample of an additional 300 PGs and controls to perform statistically powerful genetic tests to confirm and extend our preliminary findings. We found two variants of the D4 receptor gene to be associated with PG and that emotional vulnerability and impulsivity account for a large portion of the variance in PG. In the future, we will use the data from this study in combination with our dataset of ADHD patients to gain insight into the overlap of genetic risk factors in PG and ADHD, and also comorbid disorders in PG, such as depression and substance abuse.

Key words: pathological gambling, problem gambling, candidate gene studies, dopamine.

Introduction

Background Gambling is the act of betting something of value (usually money) on an uncertain outcome in the hope of a winning. The Diagnostic and Statistical Manual of Mental Disorders (A.P.A., 2000) defines pathological gambling (PG) as a persistent and recurrent maladaptive gambling behaviour that consequently disrupts personal, family, or vocational pursuits. An estimated 86% of the adult US population has gambled at least once in their lives (N.O.R.C, 1999). Most people who gamble do not develop a gambling problem; however, a fraction of these individuals will gradually escalate to larger bets and greater risks. A meta- analysis of 119 PG prevalence studies conducted in the United States and Canada reported the lifetime prevalence of PG in adults at 1.6%, with an additional 3.9% having a sub-clinical level of gambling problems, or problem gambling (PrG) (Shaffer, Hall, & Vander Bilt, 1999). A growing body of evidence is demonstrating that there is a significant genetic susceptibility to PG. When exposed to the same environmental gambling opportunities, only some individuals will develop symptoms of PG, supporting the presence of a biological susceptibility in developing the disorder in combination with environmental components (Murray, 1993). Family studies are one of the most common strategies to investigate the inheritance of complex traits or disorders, through the assessment of blood relatives of an affected individual (proband). The information is obtained either by asking the proband about their family members (family-history method) or through direct interview of the relatives (family-study method). This study design provides information on the familial segregation of the disorder, which can be the result of not only genetic, but also environmental factors. In order to determine whether a non-Mendelian disorder or trait runs in families, it is necessary to quantify the risk of a proband’s relatives to be affected by the disease compared with the population risk of the disease. This measure can be calculated for all relatives and values will be closer to one in distant relatives (Risch, 1990). Another approach to family studies is to compare the frequency of the disorder or trait under investigation and comorbid disorders in relatives of the proband and compare the obtained data to relatives of control subjects. These studies are usually conducted on first-degree relatives of the proband and comprise the majority of family studies conducted on pathological gambling. Twin studies can help elucidate the extent to which genetic factors are responsible for the development of a certain trait or disorder (phenotype). Monozygotic (MZ) or identical twins share 100% of their genes because they are derived from the same zygote. Dizygotic (DZ) or fraternal twins share, on average, 50% of their genes as do any siblings. Thus, when genetic factors account for most of the variance of a phenotype, MZ twins will present a significantly higher concordance rate compared to DZ twins. This study design assumes that twins raised in the same setting are exposed for the most part to the same environment (shared family environment). Both MZ and DZ twins are raised during the same time period, thus sharing experiences that are related to a specific time frame in the family’s history, which might not be possible for non-twins. However, environmental exposure can be different even in twins raised together, for example for experiences with different friends and the type of relationship each twin develops with the parents. This type of environmental exposure is usually referred to as

6 nonshared or unique environmental experiences. Published twin and family studies on PG are reviewed below. Family Studies in Pathological Gambling Initial evidence towards family aggregation of pathological gambling came from the finding that the risk of pathological gambling was associated with parental gambling (Jacobs, 1989; H. R. Lesieur & Blume, 1990; H. R. Lesieur et al., 1991; H. R. Lesieur & Heineman, 1988), and that children of pathological gamblers were more vulnerable to develop dysfunctional behaviors (Jacobs, 1989; H. R. R. Lesieur, J., 1989). Gambino and colleagues (Gambino, 1993) investigated the frequency of perceived addictive problems for parents and grandparents in a sample of 93 United States Veterans Affairs population of drug and alcohol abusers. Gambling behavior was assessed through the South Oaks Gambling Screen in which scores of 3 or 4 indicate a potential pathological gambler or a problem gambler and scores >5 indicate a potential pathological gambler (H. R. Lesieur & Blume, 1987). Through the South Oaks Gambling Screen scores 17.3% (n=16) of the subjects were considered as probable pathological gamblers and 14% (n=13) as potential pathological gamblers. Subjects who reported gambling problems among their parents had a three times higher liability of scoring as a probable pathological gambler. Interestingly, those who perceived their grandparents as problem gamblers were at 12 times the risk of being a problem gambler compared to those who did not report gambling problems among their grandparents. Some studies have investigated the exposure to gambling in families. Gupta and Derevensky (Gupta & Derevensky, 1997) assessed a sample of 477 children 9–14 years of age who reported gambling on a regular basis, of whom 86% reported gambling with their relatives. Although there is no evaluation regarding problem or pathological gambling among these youths, this finding points to the importance of family environment in the development of a gambling habit. In a subsequent study (Gupta & Derevensky, 1998), these authors evaluated 817 high-school students among which 4.7% were diagnosed as pathological gamblers. The adolescents who were diagnosed with pathological gambling were also more likely to report having parents with a gambling problem, a finding that has been reported in other samples of adolescents and young adults (Derevensky & Gupta, 2000; H. R. Lesieur & Blume, 1987; H. R. Lesieur et al., 1991). Other studies evaluating adults with substance abuse or dependence (Daghestani, Elenz, & Crayton, 1996; H. R. Lesieur & Blume, 1990; H. R. Lesieur, Blume, & Zoppa, 1986; H. R. Lesieur & Heineman, 1988), pathological gamblers (Ohtsuka, Bruton, DeLuca, & Borg, 1997; Volberg & Steadman, 1988), and prisoners (Walters & Contri, 1998) also reported an association of pathological gambling with parental gambling. These results add the familial factor to previous findings that have reported an association of adult problem gambling with exposure to gambling during childhood (Habra & Lee, 1995). The results of a recent investigation on a community sample of 938 adolescents (496 females, 442 males) and their parents showed that the occurrence of gambling problems in adolescents was related only to the father’s severity of problem gambling (Vachon, Vitaro, Wanner, & Tremblay, 2004). Pathological gamblers also reported higher rates of pathological gambling among their first-degree relatives. The assessment of 14 pathological gamblers through the Family History Research Diagnostic Criteria showed that at least 9% of their relatives had a gambling disorder (Black, Moyer, & Schlosser, 2003). The comparison group was not assessed for gambling behavior, but the results also show a higher frequency of pathological gambling in first degree relatives of pathological gamblers compared to the population prevalence (Shaffer & Hall, 2001).

In a larger sample of 31 pathological gamblers and 193 first-degree relatives, using both the family-history and the family-interview method, Black and colleagues (Black, Monahan, Temkit, & Shaw, 2006) reported a frequency of 8.3% for pathological gambling and 12.4% for any gambling disorder among first-degree relatives of pathological gamblers, which was significantly higher when compared with first-degree relatives of a control group (2.1% for pathological gambling and 3.5% for any gambling disorder). This is, to our knowledge, the only family study in pathological gambling with a similar male-to-female ratio sample (16 men, 15 women) and a gender-matched control group. Although the sample sizes in these family studies are relatively small, the consistency of the reports provides fairly solid support for the importance of heritable (genetic) factors in PG.

Twin Studies in Pathological Gambling The largest twin investigation in PG derives form the analysis of 3,359 male twin pairs of the Vietnam Era Twin Registry. Eisen and colleagues (Eisen et al., 1998) found that inherited factors explained between 35% and 54% of the liability for developing any symptom of PG. For example, they reported that the presence of the symptom “gambling larger amounts than intended” was composed by 55% of genetic factors and 45% of environmental factors, thus, both genetic and environmental factors contributed to this symptom with genetic factors accounting for 55% of the variance. Also, inherited factors explained 54% of the liability for the report of two or more symptoms of PG. The lifetime rates of PG among MZ and DZ co-twins of pathological gamblers were 22.6% and 9.8%, respectively and heritability estimate for the diagnosis of PG was 46%. Since heritability estimates represent a population mean, groups of individuals in the population can have increased (or decreased) biological vulnerability for PG when exposed to gambling opportunities, depending on their psychiatric comorbidity profile, unique and shared genetic and environmental factors. Family studies can provide important information regarding questions about continuity of a diagnosis by examining the risk of a particular disorder among the relatives of individuals diagnosed with either the disorder itself or with sub-clinical forms of the same condition. Thus, this sample was subsequently analyzed in order to investigate the causes of the high comorbidity between PG and alcohol dependence, and the hypothesis of a continuity model for PG (Slutske et al., 2000). Results showed that the risk of PG was significantly higher among DZ and MZ co- twins of subjects with subclinical PG (one to three Diagnostic and Statistical Manual of Mental Disorders, Third Edition-Revised PG symptoms) compared with co-twins of men with no PG symptoms. These results were consistent with the hypothesis that the differences between problem and PG are solely quantitative, meaning that these disorders represent a continuum of the same phenotype, therefore likely sharing the same risk factors. Men with subclinical PG were at a significantly higher risk for alcohol dependence when compared with men without PG symptoms. Also, the risk of alcohol dependence was higher in DZ and MZ co-twins of subjects with subclinical PG compared with those with no PG symptoms. The authors suggested that there is at least one genetic locus that increases the susceptibility for both PG and alcohol dependence. The relationship between gambling and other psychiatric disorders were also investigated in this large sample. Slutske and colleagues (Slutske et al., 2001) investigated the association of PG with antisocial behavior disorders (antisocial personality disorder, conduct disorder, and

8 adult antisocial behavior) and the results suggest that the comorbidity of antisocial disorders with PG is in part a consequence to their sharing a common genetic vulnerability. The results of an investigation on environmental and genetic contributions to the comorbidity of PG and major depressive disorder in the Vietnam Era Twin registry sample suggest that overlapping genetic factors are strong determinants of the correlation between these disorders (Potenza, Xian, Shah, Scherrer, & Eisen, 2005). These findings support the results of previous studies in smaller samples regarding a higher comorbidity rate of alcohol dependence, substance addiction, and mood disorders among first-degree relatives of pathological gamblers (Black, Monahan, Temkit, & Shaw, 2006; Black, Moyer, & Schlosser, 2003; Scherrer et al., 2005). Another twin study was conducted in a sample of 92 male and 63 female MZ and DZ twins who were assessed for overall gambling behavior during the year prior to the study (Winters & Rich, 1999). The authors found that genetic factors were significantly responsible for “high action” gambling, such as slot machines and roulette, in male twins. No significant genetic influence was found among males for “low action” games (e.g., betting on sports events or on games of skill), and no significant genetic influence among the female MZ and DZ twin pairs for both types of games. The results should be viewed cautiously, however, due to the small sample size of the twins, and the limited assessment of the gambling behavior (i.e., neither problem nor pathological gambling were assessed in the sample). The authors do not report the rate of problem or PG and do not assess any PG symptoms in the sample. In spite of the studies that have assessed the heritability of PG, little is known regarding the mode of transmission of the disorder; however, it is likely the mode of transmission of the disorder is similar to other major common psychiatric conditions, such as schizophrenia, mood disorders, and ADHD. Thus, PG is considered a “complex disorder” from a genetic perspective.

Neuropathways Associated with Pathological Gambling Neurochemistry research has indicated that the monoaminergic system may mediate PG behaviour. Initial studies have reported alterations in norepinephrine (Roy et al., 1988) and dopamine metabolites in the cerebro-spinal fluid of PG subjects (Bergh, Eklund, Sodersten, & Nordin, 1997). Also, both pathological gamblers and subjects with alcohol use disorders presented lower levels of a serotonin metabolite (5-hidroyindole acetic acid) in their cerebro- spinal fluid (Potenza, 2001). There has been consistent findings implicating the brain’s dopamine pathway in the pathophysiology of both substance and behavioural addictions (Chambers & Potenza, 2003; Chambers, Taylor, & Potenza, 2003; Potenza, 2001). Priming is a core phenomenon in substance addiction and results from the release of dopamine in the nucleus accumbens, which is located in the brain’s reward system. Similarly, gambling has been shown to produce priming-like effects in PG subjects (Zack & Poulos, 2004). Neuroimaging studies have found that the brain’s reward system is involved in PG. Pathological gamblers presented decreased ventral strialtal activation compared to controls while performing a simulated gambling task, and activation of this region had an inverse correlation with gambling severity on pathological gamblers (Reuter et al., 2005). In summary, neurochemical and neuroimaging studies of PG are converging to indicate altered monoaminergic transmission in PG subjects.

Review of Molecular Genetic Studies of PG Complex disorders have an oligogenic or polygenic genetic component with a non- Mendelian mode of inheritance, where various genes, each contributing a portion to the disease, may act together (e.g., through interaction, additive, or multiplicative effects) and interact with environmental factors. A more effective strategy for complex disorders, such as PG, is the use of association studies with candidate genes. Candidate genes are chosen from brain systems that are thought to be relevant to the etiology of the disorder. In the candidate gene approach, investigators can use at least two approaches: 1) analysis of variance of a continuous trait versus genetic variants among a group of cases (controls can or cannot be used), and/or 2) allele frequencies of the candidate gene of affected individuals compared to the allele frequencies of unaffected unrelated individuals. Both strategies were used to analyze data from our sample. Molecular genetic investigations in PG are at a very early stage, especially when compared to genetic investigations in other psychiatric disorders. Table 1 summarizes results from published studies on the molecular genetics of PG.

Table 1: Molecular genetic association studies in pathological gambling: Authors Sample Polymorphisms Results Subjects’ Controls Investigated Diagnosis Comings et al. (1996) 222 PG 714 DRD2 Taq I A Association with allele T. (Comings et al., 1996) Perez de Castro et al. 68 PG - 47 68 ethnically, DRD4 (exon III) 7-repeat allele associated with PG in (1997) (Perez de Castro, males and gender, and age females. Ibanez, Torres, Saiz-Ruiz, 21 females matched. & Fernandez-Piqueras, 1997) Comings et al. (1997) 163* 124 DRD1 DdeI Association with allele 1. (Comings et al., 1997) 186* 138 DRD2 Taq I A Association with allele T. Perez de Castro et al. 68** PG 68** 5HTT-LPR Short allele associated with PG in (1999) (Perez de Castro, males. Ibanez, Saiz-Ruiz, & Fernandez-Piqueras, 1999) Comings et al. (1999) 165* PG 124 DRD4 (exon III) 5-8 repeat and 7 repeat alleles (Comings, Gonzalez, Wu, associated with PG. Gade et al., 1999) Ibañez et al. (1999) 68** PG 68** TH (intron 1) No association. (Ibanez, Perez de Castro, Fernandez-Piqueras, & Saiz-Ruiz, 1999) Ibañez et al. (2000) 68** PG 68** MAO-A (intron 1) 4-repeat allele associated with PG in (Ibanez, de Castro, males. Fernandez-Piqueras, MAO-A (promoter) 3 repeat allele associated with PG in Blanco, & Saiz-Ruiz, males. 2000) MAO-B (intron II) No association.

Comings et al. (2001) 139 PG 139 ethnically, 31 genes involved in Genes in dopaminergic, (Comings et al., 2001) gender, and age dopamine, serotonin, noradrenergic and glutamatergic matched. noradrenaline, and GABA neurotransmitter systems accounted neurotransmitters. for < 2% of the variance in PG. Perez de Castro et al. 68** PG 68** MAO-A (promoter) 3-repeat allele associated with more (2002) (Perez de Castro, severe forms of PG in males. Ibanez, Saiz-Ruiz, & Fernandez-Piqueras, 2002) da Silva Lobo et al. 140 PG 140 non-PG full DRD1 (-800 T/C); DRD2 Association with DRD1 allele T. (2007) (da Silva Lobo et siblings (TaqI), DRD3 (Serine al., 2007) /Glycine substitution), DRD4 (exon III VNTR), DRD5 (CA repeat) * Derived from the same sample as in Comings et al. (1996). ** Same sample as in Perez de Castro et al. (1997).

The involvement of the brain’s reward system in addictions made dopamine system genes the primary targets of candidate gene studies on PG. Initial studies reported associations of PG with allele T of the dopamine D2 receptor gene (DRD2) TaqIA polymorphism (Comings et al., 1997; Comings et al., 1996), in line with the hypothesis that DRD2 was involved in the “Reward Deficiency Syndrome”(Blum et al., 1995; Comings & Blum, 2000). These associations are more difficult to interpret given the recent findings that the D2 TaqIA marker is now known to be outside the DRD2 gene, located in a neighbor gene to DRD2 called the ankyrin repeat and kinase domain containing-1 gene (ANKK1)(Neville, Johnstone, & Walton, 2004). Investigations in substance addictions have yielded conflicting results regarding associations between alcohol and nicotine addiction with the DRD2/ANKK1 TaqIA polymorphism. Some studies have found associations with neighbouring polymorphisms on ANKK1 and on other genes that flank the TaqIA site(Dick et al., 2007; Gelernter et al., 2006; Yang et al., 2007) and one study has suggested that an ANKK1 polymorphism (rs2734849, non-synonymous G to A substitution) may indirectly affect D2 receptor density(Huang et al., 2008). Findings on ADHD (Bobb et al., 2005; Misener et al., 2004), nicotine and alcohol dependence(Huang et al., 2008; Limosin et al., 2005; Sander et al., 1995; Tanabe et al., 2007), and PG(Comings et al., 1997; da Silva Lobo et al., 2007), suggest that polymorphisms in the promoter region of DRD1 may play a role in the neurobiology of addictions and disorders associated decreased impulse control. Studies using animal models for addictive behaviors found evidence that blockade of the D1 receptor prevents instrumental learning, Pavlovian conditioned learning and place preference developed by rats after ethanol, cocaine and food consumption(Cetin, Freudenberg, Fuchtemeier, & Koch, 2004; Dalley et al., 2005; Kelley, 2004; Nazarian, Russo, Festa, Kraish, & Quinones-Jenab, 2004; Sanchez, Bailie, Wu, Li, & Sorg, 2003; Self, 2004). Two studies have reported associations between DRD1 polimorphisms with PG (da Silva Lobo et al., 2007) and with comorbid PG and alcohol dependence (Comings et al., 1997). Even though the reported associations are with different polymorphisms, it is important to consider that DRD1 is a small gene, in which these two polymorphisms are in a high degree of linkage disequilibrium, which means that these polymorphisms have been transmitted together throughout evolution with little interference from recombination processes. Together, results from these studies make it reasonable to hypothesize that D1 receptor function arising from variation in expression of the DRD1 gene may affect vulnerability to addictions, possibly through and increased vulnerability to environmental cues. A polymorphism in exon III of the dopamine D4 receptor gene (DRD4) exon III variable- number-of-tandem-repeats (VNTR) has been reported to encode a receptor with lower affinity for dopamine(Asghari et al., 1995; Jovanovic, Guan, & Van Tol, 1999; Van Tol et al., 1992), and to be associated with impulsive personality traits (Benjamin et al., 1996; Ebstein et al., 1996; Strobel, Wehr, Michel, & Brocke, 1999). The association of the DRD4 exon III VNTR with PG was investigated in two studies (Comings, Gonzalez, Wu, Gade et al., 1999; Perez de Castro, Ibanez, Torres, Saiz-Ruiz, & Fernandez-Piqueras, 1997), and associations with the 7-repeat allele and with the overall long forms of the polymorphism (5–8 repeats) (Comings, Gonzalez, Wu, Gade et al., 1999) were reported. It should be noted that in one study (Perez de Castro, Ibanez, Torres, Saiz-Ruiz, & Fernandez-Piqueras, 1997) this polymorphism was reported to be associated only with the vulnerability to PG in females. However, the small number of females in the sample has to be considered as a limitation in the interpretation of a sex-effect on DRD4 and PG.

Deficient impulse control and impulsive personality features also have suggested the involvement of serotonergic and noradrenergic pathways in PG (Blanco, Orensanz-Munoz, Blanco-Jerez, & Saiz-Ruiz, 1996; DeCaria et al., 1996; Roy et al., 1988; Zuckerman, 1988). A variant in the promoter region of the serotonin transporter gene (5HTTLPR) was found to be significantly associated with PG in males(Perez de Castro, Ibanez, Saiz-Ruiz, & Fernandez- Piqueras, 1999). The long variant of this polymorphism is associated with increased promoter activity and thus higher production of the serotonin transporter protein. Subsequent investigations in this same sample revealed an association of the more severe forms of PG in males with a 3-repeat allele in the promoter region of the monoamine oxidase A (MAO-A) gene VNTR and with a polymorphism in the first intron of the same gene(Ibanez, de Castro, Fernandez-Piqueras, Blanco, & Saiz-Ruiz, 2000; Perez de Castro, Ibanez, Saiz-Ruiz, & Fernandez-Piqueras, 2002). The MAO-A gene is located on the X chromosome, thus, there is a greater likelihood that sex differences in its expression may occur. Comings and colleagues(Comings et al., 2001) analyzed the effect of 31 genes involved in dopamine, serotonin, norepinephrine, and γ-aminobutyric acid pathways in 139 pathological gamblers and 139 controls. The most significant associations were found with the D2 and D4 receptors, the dopamine transporter, tryptophan hydroxylase, the α-2C adrenergic receptor, the glutamate receptor subunit 1, and the presenilin 1 genes. These results indicate that genes influencing a range of brain functions may play an additive role as risk factors for PG. One of the main limitations of these previous studies is the sample size, which increases the potential for both type I and type II errors.

Other monoaminergic system genes and PG DRD3. Sokoloff et al. (Sokoloff, Giros, Martres, Bouthenet, & Schwartz, 1990) characterized the gene encoding of DRD3. It is localized in the limbic system, with strong expression in the nucleus accumbens, an area of the brain involved in the reward system. Mice deficient in D3 have been shown to exhibit hyperactivity (Accili et al., 1996). Pilla et al. (Pilla et al., 1999) showed BP897, a highly selective D3 receptor agonist, was able to inhibit cocaine- seeking behaviour induced by drug-associated cues. D3 expression was increased in a nicotine addiction model (Le Foll, Diaz, & Sokoloff, 2003). Thus, there is a strong rationale for a role of D3 ligands as therapeutic tools in the treatment of addictive and impulsive behaviour. The DRD3 gene has been tested previously in association studies of dependence on cocaine (Ballon et al., 2007; Comings, Gonzalez, Wu, Saucier et al., 1999), alcohol (Lee & Ryu, 2002; Limosin et al., 2005; Sander et al., 1995; Szczepankiewicz et al., 2007; Wiesbeck et al., 2006), and heroin (Li et al., 2002), but results were mixed. Studies of DRD3 in ADHD yielded negative results (Ballon et al., 2007; Barr et al., 2000; Kopeckova et al., 2008; Muglia, Jain, & Kennedy, 2002; Payton et al., 2001; Retz, Rosler, Supprian, Retz-Junginger, & Thome, 2003). The discrepant findings could be due to the use of only one polymorphism (Ser9Gly, rs6280) within DRD3. Also, other regions of the gene were not investigated extensively. DRD3 promoter polymorphisms were recently studied in schizophrenia with positive results, but their functional relevance has not been assessed yet (Anney et al., 2002; Sivagnanasundaram et al., 2000; Staddon et al., 2005). DAT1. The DA transporter (DAT1), which acts to take released DA back up into storage in presynaptic terminals, has been implicated in various psychiatric disorders, including disorders with impulse control dysregulation such as ADHD (Bellgrove & Mattingley, 2008;

Elia & Devoto, 2007; Kuntsi, McLoughlin, & Asherson, 2006). Extracellular DA in the neuronal synapse seems to play a key role in frontal cortex activity, also a part of the brain’s reward system (Kirley et al., 2002). Conflicting lines of evidence have been reported on the association between the 480-bp (10-repeat allele) polymorphism of the 40-bp DAT1 VNTR and ADHD (Kirley et al., 2002). Currently, no studies have investigated the role of DAT1 in PG.

Research Hypotheses: 1. General Hypothesis: Polymorphisms in dopamine system genes will be associated with pathological gambling when compared to control subjects. 2. Specific hypothesis: • Low functioning alleles (2 and 7-repeat) of the D4 gene (DRD4) will be associated with PG when compared to control subjects and with higher SOGS scores. • DRD2 (TaqIA) and DRD3 polymorphisms will be associated to PG when compared to control subjects.

Methods Participants and Questionnaires Through two previous studies funded by the Ontario Problem Gambling Research Centre (OPGRC) entitled Identification of genetic risk factors for pathological gambling, and Special Problems in gambling: Attention Deficit Hyperactivity Disorder (ADHD) and pathways to problem gambling we collected DNAs from 145 subjects with different degrees of gambling severity. The following instruments were used in the present study: 1. Winning Experiences Questionnaire (Turner, unpublished) 2. Gambling Behavior Questionnaire (Turner, unpublished) 3. Random Events Knowledge Test (Turner, unpublished) 4. Ways of Coping Questionnaire (Folkman & Lazarus, 1988a, , 1988b) 5. Gambling Strategy Questionnaire (Turner, unpublished) 6. Inventory of Gambling Situations (Littman-Sharp & Stirpe, 1997) 7. Impaired Control Scale (Heather, Tebbutt, Mattick, & Zamir, 1993) 8. Problem Solving Inventory (Heppner, 1988) 9. Safe at Play Risk-Quiz (available through http://www.problemgambling.com/wall.html) 10. Zung Self Rating Depression Scale (Zung, 1965) 11. Temperament and Character Inventory (Cloninger, Svrakic, & Przybeck, 1993) 12. Barrat Impulsiveness Scale (Patton, Stanford, & Barratt, 1995) 13. Structured Clinical Interview Diagnosis in Psychiatry (A.P.A., 1994) 14. Canadian Problem Gambling Index (Smith & Wynne, 2002) 15. Gambling Information Questionnaire (Turner, unpublished) 16. Gambling Cognition Questionnaire (Toneatto, 1999) Subjects in the first two projects completed a thorough questionnaire evaluating the level of severity for gambling, including the South Oaks Gambling Screen (SOGS), a self-report

14 version of the DSM-IV for PG, and responses to a set of questions about harmful consequences of gambling. A number of other cognitive variables were tested in this sample, such as the Random Events Knowledge Test – REKT (Turner, unpublished) and the Gambling Cognition Questionnaire (GCQ), developed by Toneatto et al. (Toneatto, 1999) to measure erroneous beliefs about gambling found in problem gamblers. In the 3 years of the current study, we collected a total of 600 subjects with varying problem gambling severity and non-pathological gamblers. The groups were categorized using SOGS scores as follows: 0 = control; 1-2 = at-risk; 3-4 = problem gambler; and 5-20 = pathological gambler (PG). In addition to this sample, we have analyzed a Brazilian sample of 140 discordant sib- pairs (one PG and one non-PG sibling), which were obtained through our collaboration with Dr. Daniela Lobo while at the University of São Paulo, São Paulo, Brazil. Brazilian Caucasians, particularly in São Paulo, have a relative high contribution from the European gene pool (Callegari-Jacques et al., 2003) making them reasonably comparable to the Canadian sample. Genotypes in both samples were in HWE, which was also maintained in the pooled sample. Therefore, it is unlikely that our results are due to population stratification. The Brazilian sample was used specifically to validate our findings regarding DRD4 in the Toronto sample.

Laboratory Methods Blood samples were collected from subjects and DNA was extracted from whole blood following standard high salt procedures. The regions of interest within the candidate genes were amplified using polymerase chain reaction (PCR) techniques with primers and conditions previously published. The PCR products for the following candidate genes were visualized via gel electrophoresis performed in agarose gels prepared with ethidium bromide and 1X TBE (Tris, boric acid, EDTA): DRD4, DAT, MAOA, and 5HTT. Upon the introduction of the Sequence Detection System ABI 7500 (Applied Biosystem Inc., Canada), COMT, Brain Derived Neurotrophic Factor (BDNF), DRD1, DRD2, and DRD3 polymorphic regions were amplified and genotyped using TaqMan® SNP Genotyping Assays (Applied Biosystems Inc, USA). The allelic assignment was performed using the ABI Prism 7500 Sequence Detection System (SDS) SoftwareTM (Applied Biosystems). The DRD5 and TH PCR products were separated using an ABI 3130-Avant Automated Genetic Analyzer (Applied Biosystem Inc., Canada). Allele assignment was performed using GeneMapper 3.0TM. The subjects were genotyped blind to their affection status.

Data Analysis Demographic and clinical data of the Toronto sample were analyzed using the chi-square test for dichotomous response variables, and the independent samples t-test for continuous response variables. Demographic analysis and characteristics of the Brazilian sample have been previously described (da Silva Lobo et al., 2007).

Analysis of Correlates of Pathological Gambling

Correlation tests were performed to verify which variables were significantly associated with PG and with PG severity. Subsequently, a factor analysis was performed in order to verify whether the variables correlated with PG would load on different factors. These factors were then extracted as factor scores and used in a regression analysis, which showed how this factors contributed to the variance in PG.

Candidate gene Analysis Sample power calculations for the quantitative dependent variable (lifetime SOGS score) and the case-control analysis were performed through QUANTO software (Gauderman & Morrison, 2006). Overall, single marker (SNP, polymorphism) association analysis for PG was performed using chi-square tests through the software UNPHASED version 3.0.10 (Dudbridge, 2003). Since it was possible to genotype 9 SNPs for the DRD3, we performed both a single marker and a haplotype analysis through the software Haploview (http://www.broad.mit.edu/node/443 ). Haplotypes are blocks inside the gene that, throughout evolution, have been transmitted together and were not affected by recombination. These haplotype blocks can be formed by 2 or more SNPs and we investigated whether combinations of the SNPs in each block would be associated with PG. Our specific hypothesis referred to the association of DRD4 with both the diagnosis of PG and with severity of PG as measured by SOGS scores. Statistical analysis for the case-control design (PG diagnosis x DRD4) was performed through UNPHASED version 3.0.10 (Dudbridge, 2003), using the extended pedigree-disequilibrium test, which allows for the combination of families and singletons. The allele main-effects analysis for the DRD4 (exon III) in the Brazilian sample (da Silva Lobo et al., 2007) was performed with a previous version of UNPHASED (Dudbridge, 2003). Statistical analysis for the quantitative trait design (SOGS scores x DRD4) was performed using univariate analysis of variance (ANOVA) followed by a post-hoc Tukey test. These analyses were carried out on SPSS version 14 (SPSS Inc, Chicago, IL). The traditional Bonferonni approach is considered to be excessively conservative for candidate gene studies (Nicodemus, Liu, Chase, Tsai, & Fallin, 2005). Since a number of tests were performed in this dataset, we corrected for multiple testing using permutation tests (Potter, 2006). Permutation tests provide an empirical p-value derived from multiple permutations of the information from the dataset, thus providing an estimate of how likely the association occurs by chance and is considered the gold standard for multiple testing correction in genetic association studies (Nicodemus, Liu, Chase, Tsai, & Fallin, 2005). All the p-values reported in the Results section have been corrected for 1,000 permutations.

Results Sample Categorization Over the 3 years of this project, 600 subjects consented to participate and were recruited. Thirty of the subjects were unable to complete the interview or the questionnaires (Table 2).

Table 2. Summary of sample categorization determined by SOGS scores SOGS score Category Males Females Total 0 control 46 71 117 1-2 At-risk 43 54 97 3-4 problem gambler 24 15 39 5-20 pathological gambler 222 95 319 Undetermined* Excluded 30 TOTAL 335 235 600 Note. * Unable to categorize due to incomplete questionnaires.

The Canadian sample had a mean (±SD) age of 41.7 years (±11.5) for PG subjects, and 36 years (±12.8) for controls (p= 0.02). The sib-pairs were matched for age (PG mean age 40 ±8.9 years; siblings mean age 40 ±9.8 years). The combined sample had a mean (±SD) age of 41.5 years (±11) for PG subjects, and 36 years (±12.5) for controls (p<0.001). Pathological gamblers reported a mean of 11.5 ± 4 symptoms on the SOGS in the Canadian sample, and a mean of 15 ±3.1 symptoms in the Brazilian sample.

Correlates of Pathological Gambling In order to simplify the correlation and regression analysis, we combined the DSM-IV, SOGS, and CPGI into a single aggregate variable. As shown in Table 3, the pattern of correlations was similar using the DSM-IV, SOGS, Lifetime SOGS, or the aggregate problem gambling scores. Several of the psychological variables were significantly associated with PG. PG was significantly correlated with depression, Zung anxiety scores, impulsiveness, and boredom susceptibility, as well as with reported use of gambling as an escape and avoidance stress avoidance mechanism (Table 2). Table 3. Correlates of PG , Spearman's rank-order correlation test. Correlates Aggregate DSM-IV total SOGS 1-yr SOGS life Impulsiveness .314** .326** .289** .282** Boredom susceptibility .177** .193** .138** .174** Depression .489** .510** .461** .418** Zung anxiety .495** .510** .480** .413** WCQ escape and avoidance .466** .474** .442** .412** Timing of wins First gamble ever .150** .101* .157** .165** Early days after learning how .219** .177** .204** .231** Every now and then .081 .049 .81* .94* First time ever -.012 -.02 -.008 -.008 Log size of first win .294** .252** .299** .268** Random events knowledge -.270** -.240** -.290** -.210** Note. ** p < .01; * p < .05

Knowledge of random events was significantly negatively associated with PG. Each of these effects replicates our previous findings reported in Turner, 2006 (Turner, 2006) and adds additional evidence in support of the Blaszczynski and Nower (Blaszczynski & Nower, 2002) pathways model. Consistent with a behavioral model of gambling addiction, severity of problematic gambling is correlated to the experience of winning (rho = 31, p < .001) and, in particular, winning the first time (rho = .22, p < .01) or winning early in one’s gambling career (rho = .25, p < .001). These effects are small, suggesting that winning only partially explains the development of a gambling problem. Interestingly, winning every now and then (rho = .09, ns) and winning recently (rho = .00, ns) were unrelated to problem gambling. This suggests that early wins rather than wins in general may lead to pathological gambling. In addition we also found that the size of a person’s first win was correlated with severity of problematic gambling (rho = .22, p < .01). Interestingly, the strongest difference was not in the amount or timing of the win, but in the psychological effect of the win. There were strong correlations between problematic gambling and the feeling of wanting to gamble more after the win, (rho= .56, p < .01), the belief that one could beat the odds, (rho = .45, p < .01), and a modest correlation with the belief that one could win again, (rho = .21, p < .05). Severity of PG was also correlated with all of the subscales of the TCI (Table 4). The strongest effects were for self-control, rho = -.50, p < .001, cooperativeness, rho = -.39, p < .001, and damage avoidance, rho = .35, p < .001.

Table 4. Correlations between TCI higher order dimensions and PG severity, Spearman's rank- order correlation test.

Temperament and character scales rho p TCINS curiosity behaviour .35 .000001 TCIHA harm avoidance .24 .000001 TCIRD reward dependency -.20 .000001 TCIPR persistence ability -.12 .002048 TCISD self-control capacity -.50 .000001 TCICO cooperativity -.39 .000001 TCIST self-transcendency .14 .000504

We pooled the variables that were correlated with PG and entered 18 variables into factor analysis with the aim of reducing the collection of predictors into a set of basic themes. Five eigenvalues exceeded 1.0. We tried various extraction routines and determined that all five eigenvalues created factor scores that correlated with PG severity. These factors were extracted as factor scores and used in a regression analysis. Table 5 lists the regression weights of factor scores onto PG severity. Emotional Vulnerability was by far the strongest effect accounting for 35.4% of the variance of pathological gambling. Based on this analysis it would appear that Pathological gambling is more of an emotional disorder than an impulse control disorder, however it should be pointed out that 3 of the impulse control variables were also loaded on the emotional vulnerability pathway which suggested that the emotional pathway is a combination of emotional vulnerability and impulsivity.

Table 5. Regression weights of factors scores onto the severity of PG Factors Beta p Emotional vulnerability .38 .001 Impulsivity .28 .001 Cooperativeness/dependency -.15 .001 Early wins .34 .001 Erroneous beliefs .35 .001 Overall effect R2 = .479 Controlling for ethnic background and gender R2 = .512

Results from Candidate Gene Analysis

We investigated one polymorphism on each of the following genes: DRD1, DRD2, DRD4 and DRD5 (dopamine receptor genes type 1, 2, 4, and 5 respectively), and DAT1 (dopamine transporter gene). We performed a haplotype analysis on 9 polymorphisms in the DRD3 (dopamine receptor gene type 3) and examined one polymorpfism in BDNF (brain derived neurotrophic factor), which partially controls the expression of DRD3 (Guillin et al., 2001). Association analysis of DRD1 (p 0.08), DRD2 (p 0.06) and BDNF (p 0.5) have not yielded positive results. Both single marker and haplotype analysis of DRD3 revealed no association with PG, however since this is the first haplotype analysis we performed in PG, we will demonstrate the results below.

DRD3 Results:

The 9 SNPs (polymorphisms) genotyped for the DRD3 include the following:

• DRD3 (rs6280, ser9gly) • DRD3 (rs1025398) • DRD3 (rs7611535) • DRD3 (rs167770) • DRD3 (rs2087017) • DRD3 (rs1394016) • DRD3 (rs2134655) • DRD3 (rs905568) • DRD3 (rs2399504)

The haplotype structure of the DRD3 gene, for Caucasians, is presented in Figure 1. As we can see, there are 3 blocks inside the gene: the first block includes SNPs rs2399504 and rs7611535; the second block includes SNPs rs1394016 and rs6280, and the third block includes rs2134655 and rs2087017. In this sample, chi-square tests for association and haplotype tests were negative. Results are summarized below in Tables 6 to 11.

Figure 1. DRD3 haplotype block structure (Caucasians).

Table 6. Allele main-effects association test for DRD3 markers (317 PGs, 227 controls).

Cases Controls SNP Alleles χ2 p n (%) n (%) C 388 (61) 244 (62) rs905568 .024 .8777 G 266 (39) 164 (38) G 511 (81) 121 (82) rs2399504 .1 .7516 A 351 (19) 79 (18) A 466 (74) 168 (75) rs7611535 .143 .7054 G 322 (26) 110 (25) C 310 (53) 278 (54) rs1394016 .097 .7560 T 232 (47) 200 (46) A 341 (61) 223 (60) Ser9Gly .108 .7427 G 246 (39) 168 (40) C 409 (65) 221 (66) rs167770 .039 .8432 T 283 (25) 149 (24) A, 449 (76) 143 (75) rs2134655 .009 .9228 G 325 (24) 105 (25) C 345 (54) 289 (55) rs2087017 .023 .8806 T 236 (46) 194 (45) A 392 (67) 190 (64) rs1025398 .964 .3262 G 273 (33) 151 (36)

Table 7. Haplotype association test for DRD3 markers (317 PGs, 227 controls)

Haplotype Cases Controls Haplotype blocks χ2 p frequency n (%) n (%) rs2399504 - rs7611535 AG .738 466 (73.5) 321(74.3) .086 .7695 GA .187 122(19.1) 78(18.0) .18 .6718 AA .074 47(7.4) 32(7.4) .001 .9973 rs1394016 - Ser9Gly CA .527 333(52.8) 228(52.7) .0 .9902 TG .402 255(40.4) 172(39.9) .022 .8815 CG .065 42(6.6) 28(6.4) .02 .8872 rs2134655 - rs2087017 AC .447 285(45.0) 190(44.2) .062 .8033 AT .310 195(30.8) 135(31.4) .042 .8380 GT .236 150(23.7) 101(23.5) .0020 .9629

Table 8. Allele main-effects association test for DRD3 markers in 266 male subjects (220 PGs, 46 controls)

Cases Controls SNP Alleles χ2 p n (%) n (%) C 274(63) 51(58) rs905568 .657 .4175 G 164 (37) 37(42) G 350(80) 72(84) rs2399504 .666 .4143 A 88 (20) 14(16) A 316(72) 66(75) rs7611535 .371 .5424 G 124 (28) 22(25) C 205(50) 51 (58) rs1394016 1.835 .1755 T 205(50) 37(42) A 232(60) 51(64) Ser9Gly .369 .5433 G 154 (40) 29(36) C 274(63) 62(71) rs167770 1.981 .1593 T 164 (37) 26(29) A 310(75) 65(74) rs2134655 .074 .7862 G 102 (25) 23(26) C 248(56) 44(50) rs2087017 1.201 .2730 T 192 (44) 44(50) A 274(69) 60(68) rs1025398 .015 .9035 G 124 (31) 28(32)

Table 9. Haplotype association test for DRD3 markers in 266 male subjects (220 PGs, 46 controls)

Haplotype Cases Controls Haplotype blocks χ2 p frequency n (%) n (%) rs2399504 - rs7611535 AG .723 316(71.8) 66(75.0) .371 .5424 GA .193 88(20.0) 14(15.9) .787 .3749 AA .083 36(08.2) 8(9.1) .079 .7782 rs1394016 - Ser9Gly CA .516 221(50.5) 50(56.8) 1.142 .2853 TG .405 181(41.5) 32(35.6) 1.038 .3083 CG .077 34(7.9) 6(6.4) .239 .6249

We have also performed a separate single marker and haplotype analysis for males and females in order to investigate a possible sex-effect of DRD3 in PG. There was no evidence of a sex-effect, as it is demonstrated in tables 10 and 11.

Table 10. Allele main-effects association test for DRD3 markers in 278 female subjects (95 PGs, 183 controls). Cases Controls SNP Alleles χ2 p n (%) n (%) C 114(58) 215(63) rs905568 .879 .3485 G 80(42) 127(37) G 161(83) 279(81) rs2399504 .296 .5865 A 33(17) 65(19) A 150(77) 256(74) rs7611535 .564 .4527 G 44(23) 88(26) C 105(59) 181(53) rs1394016 1.923 .1655 T 73(41) 163(47) A 109(61) 195(58) Ser9Gly .392 .5314 G 69(39) 139(42) C 135(70) 221(64) rs167770 2.034 .1538 T 57(30) 123(36) A, 139(77) 260(76) rs2134655 .094 .7590 G 41(23) 82(24) C 97(50) 192(56) rs2087017 1.878 .1705 T 97(50) 150(44) A 118(64) 213(63) rs1025398 .028 .8672 G 66(36) 123(37)

Table 11. Haplotype association test for DRD3 markers in 278 female subjects (95 PGs, 183 controls)

Haplotype Cases Controls Haplotype blocks χ2 p frequency n (%) n (%) rs2399504 - rs7611535 150 (77) 255(74) AG .753 .68 .4097 44 (23) 89(26) 33(17) 64(19) GA .180 .213 .6445 161(83) 280(71) 11(6) 24(7) AA .065 .349 .5549 183(94) 320(93) rs1394016 - Ser9Gly 112(57) 178(52) CA .539 1.813 .1781 82(43) 166(48) 73(38) 141(41) TG .398 .584 .4447 121(62) 203(59) 8(4) 22(6) CG .055 1.419 .2335 186(96) 322(94) rs2134655 - rs2087017 97(50) 149(44) AC .458 2.046 .1526 97(50) 193(56) 53(27) 111(32) AT .306 1.679 .1951 141(73) 231(68) 44(23) 81(24) GT .233 .042 .8385 150(77) 261(76)

DRD4 Gene In accordance with the molecular findings on the D4 receptor function (decreased adenylyl cyclase activation in cell lines expressing 2- or 7-repeat alleles), we separated DRD4 alleles in two groups for the main-effects test: Group 1: 3, 4, 5, 6, 8, 9, and 10-repeat alleles; Group 2: 2 or 7-repeat alleles. For the genotype main-effects test, genotypes were grouped as follows: Group 1/1 = genotypes without any 2- or 7-repeat allele; Group 1/2 = genotypes with either one 2- or one 7-repeat allele; Group 2/2 = genotypes 2/2, 2/7, or 7/7.

Case-Control Analysis for DRD4 and PG The final case control sample, with successful genotyping for this locus, included 260 PGs and 185 Controls. The estimated power to detect association is 95% for an odds ratio as low as 1.6, a risk allele frequency of 22% (20% for the 7-repeat allele and 2% for the 2-repeat allele), and an additive mode of inheritance with a 5% alpha (2-tailed p-value).

The preliminary analysis of association in the Canadian sample revealed nominal significant associations for both the 7-repeat allele and the 7/7 genotype (Table 12), but with a small number of cases and controls. When the DRD4 (exon III) alleles and genotypes were grouped, the association with PG remained significant, although limited by a small number of controls with the risk alleles/genotypes (Tables 13 and 14). Table 12. Allele and Genotype Main-Effects Test for DRD4 exon III VNTR polymorphism (260 PGs, 185 controls) *

Frequency in cases Frequency in controls Odds ratio Alleles χ2 p n (%) n (%) (95% CI) 2 repeats§ 54 (10) 35 (9) 1 .21 .65 3 repeats 17 (3) 10 (3) 1.1 (.47 – 2.56) .23 .63 4 repeats 333 (64) 261 (71) .83 (.52 – 1.31) 4.25 .04 5 repeats 6 (1) 7 (2) .56 (.17 – 1.80) .81 .37 7 repeats 105 (20) 51 (13) 1.33 (.80 – 2.23) 5.65 .02 Frequency in cases Frequency in controls Odds ratio Genotypes χ2 p n (%) n (%) (95% CI) 2/2† 3(1) 2(1) 1 – – 2/4 34(15) 31(18) .73(.1 – 4.7) .91 .34 3/4 12(5) 7(4) 1.14(.2 – 8.6) .25 .62 4/4 114(49) 90(53) .84(.1 – 5.2) .57 .45 4/7 50(22) 33(20) 1.01(.2 – 6.4) .28 .60 7/7 19(8) 7(4) 1.81(.3 – 13.2) 3.3 .07 Note. * Results are shown for allele counts ≥ 5 (omitted alleles: 6, 8, and 9 repeats). Test with 7 df for allele main-effects and with 5 df for genotype main effects test. § Reference allele. † Reference genotype.

Table 13. Allele Main-Effects Test for DRD4 exon III VNTR polymorphism, grouped according to functional evidence (260 PGs, 185 controls)

Frequency in cases Frequency in controls Odds ratio Alleles χ2 p n (%) n (%) (95% CI) Group 1§ 361(69%) 284(77%) 1 5.77 0.02 Group 2 159(31%) 86(23%) 1.5 (1.1 – 2.0) Note. §Reference allele.

Table 14. Genotype Main-Effects Test for DRD4 exon III VNTR polymorphism, grouped according to functional evidence (260 PGs, 185 controls)

Frequency in cases Frequency in controls Odds ratio Genotypes χ2 p n (%) n (%) (95% CI) Group 1/ 1§ 135(52%) 108 (58%) 1 1.86 .17 Group 1/ 2 91(35%) 68(37%) 1.1 (.7 – 1.6) .15 .70 Group 2/2 34(13%) 9(5%) 2.94(1.4 – 6.6) 11.41 .0007 Note. §Reference genotype.

SOGS and DRD4 Analysis Results: The final sample for this analysis was comprised of 439 PG, 50 problem gamblers, 103 at-risk (or non-problem) gamblers and 205 controls (797 subjects). The estimated power to detect association was 85% for an r2 of 0.01, assuming an allelic frequency of 22%, 5% alpha, population frequency of 2%, and an additive mode of inheritance. Non-PG siblings were excluded from this analysis in order to maintain a sample of unrelated subjects. Since there was a significant difference in age and sex of cases and controls, these variables were included as covariates in the final model. Table 15 shows the mean SOGS score for subjects on each of the genotype groups for DRD4. The Kruskall-Wallis test for the association between DRD4 and SOGS scores is shown on table 16. The effect of DRD4 (exon III) genotype remains significant (p= 0.003) with age and sex correction. The Tukey test (Table 17) shows that group 2/2 genotypes are significantly associated with higher SOGS scores (mean 9.52 ±6.63, group 2/2 x group 1/1 – p=0.006, d = .38; group 2/2 x group 1/2 – p=0.01, d = .38) compared to the means of the other 1/1 and 1/2 group genotypes (7.01 ±6.62 and 7.02 ±6.61, respectively. Group 1/1 x group 1/2 – p=0.9, d = .001). Finally, we investigated whether sex or age would moderate the association between SOGS and DRD4 (exon III). Age and sex were independently associated with SOGS (p <0.001 and p<0.001, respectively), but no interaction was found with DRD4 (exon III) genotypes with either age (p=0.40) or sex (p=0.42).

Table 15. SOGS scores according to genotype groups for 657 subjects (299 PGs, 50 problem gamblers, 103 at-risk gamblers, 205 controls).

DRD4 (exon III) Frequency in the total sample SOGS mean (± SD) genotypes n (%) Group 1/1 368 (56) 5.40 (± 5.96) Group 1/2 226 (34.4) 5.33 (± 5.76) Group 2/2 63 (9.6) 8.43 (± 6.51)

Table 16. Kruskal-Wallis test for 657 subjects (299 PGs, 50 problem gamblers, 103 at-risk gamblers, 205 controls), investigating the association between SOGS scores and DRD4 (exon III) genotypes grouped according to functional evidence.

Parameter N Mean rank χ2(df) p Group 1/1 368 320.26 Group 1/2 226 320.13 13.74 (2) .001 Group 2/2 63 411.88

Table 17. Post-hoc Tukey HSD test for DRD4 (exon III) genotypes associated with SOGS scores in a sample of 657 subjects (299 PGs, 50 problem gamblers, 103 at-risk gamblers, 205 controls).

DRD4 (exon III) Mean SE p 95% CI genotypes Difference Group 1/1 x Group 1/2 .07 .50 .99 -1.11 – 1.26 Group 2/2 x Group 1/2 3.10 .84 .001 .49 – 4.17 Group 2/2 x Group 1/1 3.02 .81 .001 .51 – 4.47

Since DRD4 was significantly associated with PG, we performed a regression analysis of DRD4 and PG, and a correlation analysis between the variables extracted in the factor analysis and DRD4. A regression of the high-risk genotypes of the DRD4 exon III polymorphism indicated it accounted for significant variance in problem gambling severity, by itself accounting for 1.7% of the variance of PG. When we added in the psychometrics, most of the variance overlapped with the other factors, particularly with emotional problems. Table 18 lists the correlations between the DRD4 high-risk genotypes and the various factors identified in the analysis. The effect of the high-risk DRD4 genotypes appeared to be mostly indirect, through emotional vulnerability. This suggests the genotype influences the tendency of PGs to seek out or rely on negative reinforcement (e.g., escape and avoidance, wishing one’s problems away). By itself, this genotype added 4% unique variance to the model. However, we know that this polymorphism is also associated with impulsive traits and with cognitive correlates of impulsivity, which means that its influence could be underestimated because it is currently not possible to extract its influence on other variables in the model.

Table 18. Correlation of factors with DRD4 risk genotype 2/2; 2/7; 7/7; Spearman's rank-order correlation test.

Factors r p Gambling problems .14 .001 Emotional vulnerability .11 .001 Impulsivity -.02 ns Cooperativeness/dependency -.02 ns Early wins .03 ns Erroneous beliefs .07 .06

In addition, we computed an interaction term for the high-risk DRD4 genotypes and each of the factors that accounted for significant variance of PG. The reason for conducting this analysis was to determine whether there were any interactions between gamblers’ experiences and their genetic make up. We found a significant interaction between the high-risk genotypes and experiences of wins. The combination of the risk alleles and the experience of early wins had a greater than additive effect on the severity of PG that added 6% to the total model. For people without one of the high-risk genotypes, the correlation between experiences of wins and severity of problematic gambling was rho = .32, p < .001. However, for people with one of the high risk genotypes, the

26 correlation between experiences of wins and severity of problematic gambling was rho = .50, p < .001. Other genes examined in this study (DAT1, DRD2, DRD3, DRD5) have so far not accounted for additional variance of PG; however, we have not yet examined whether combinations of these genes are significantly associated with PG severity or whether any of these genotypes interact with the psychological variables. It is also important to mention that, according to the initial proposal, only one polymorphism on each gene was investigated in this study, with the exception of DRD3.

Discussion Many psychiatric disorders are associated with abnormal function of the limbic system, an area of the brain that regulates emotions, as well as reward and impulsivity. Guillin et al. (Guillin et al., 2001) reported that BDNF exerts significant control of D3 expression in the nucleus accumbens in an animal model, leading to the hypothesis that dysregulation within the BDNF-DRD3 pathway is involved in the psychiatric disorders in which deficient impulse control and abnormal reward-seeking behaviour are core features. In this study, no association was observed between DRD3 genetic polymorphisms and PG; however, given the role of DRD3 in addictive behaviours, further examination is justified. It may be the case that the effect of DRD3 is dependent on interaction with one or more other genes. BDNF plays a key role in neuronal survival in the developing and mature central nervous system. Like DRD3, BDNF is also expressed in the accumbens and other reward areas of the brain. Disorders, such as ADHD, nicotine dependence, problem gambling, bulimia nervosa/binge eating, and OCD, share common problems with impulsivity and/or compulsions. Patients may also have combinations of these symptoms. These illnesses are highly heritable; specifically, 80% in ADHD, 50% in smoking initiation, 70% in persistent smoking, 60% in problem gambling, 28-83% in bulimia nervosa, and 30-40% in OCD. Genetically, all of these disorders have been associated with the DA system, and the majority has been linked to BDNF. In addition, drugs of abuse have been shown to modulate the expression of BDNF together with DRD3, supporting the hypothesis that BDNF gene may be involved in dependence, at least partly, through DRD3 (Le Foll, Diaz, & Sokoloff, 2005). Genome-scan association studies using markers that flank the BDNF locus support the involvement of this gene in addiction vulnerability (Uhl, Liu, Walther, Hess, & Naiman, 2001). This hypothesis is strengthened by the recent findings of a significant association between allele 228 of the BDNF marker ss13534842 and substance abuse in a large sample (Liu et al., 2005). In contrast, no significant results were found between the BDNF Val66Met rs6265 polymorphism and either poly-substance abuse or alcoholism. The BDNF gene was recently studied in relation to tobacco dependence, and a four- marker haplotype was found associated in Caucasian males (Beuten et al., 2005). BDNF has also been associated with ADHD and OCD, though replication studies are needed (Friedel et al., 2005; Hall, Dhilla, Charalambous, Gogos, & Karayiorgou, 2003; Kent et al., 2005). A further step would be to examine multiple SNPs (haplotypes) in BDNF and the DRD3-BDNF interaction in this sample. The DRD4 48bp repeat exon III is of particular interest because it has been linked to several impulse control disorders, including PG. We are the first genetics group to sort the DRD4 variants into low 2R, 7R versus high 4R signal transduction (Jovanovic, Guan, & Van Tol, 1999;

Van Craenenbroeck et al., 2005). We found significant association of DRD4 exon III polymorphism with PG. This could be explained by reported association of the DRD4 exon III polymorphism with novelty seeking, combined with the high level of novelty seeking associated with PG. We tested for the interaction between novelty seeking and DRD4, as well as for DRD4 and sex, in our sample; however, due to a restricted number of PG subjects, the analyses likely lacked power to detect the interaction effect. The overlap between the high-risk DRD4 genotypes and emotional problems, and the interaction between high-risk DRD4 genotypes and early wins are both consistent with the idea that PGs suffer from a dysregulation of dopaminergic transmission. Myopia for future consequences (Cavedini, Riboldi, Keller, D'Annucci, & Bellodi, 2002) may be due to both an avoidance of negative reinforcement (e.g., escape from depression, anxiety, and stressful life experiences) and an over-sensitivity to positive reinforcement (e.g., a greater rush from an early big win). Currently, however, we can only account for a small amount of the variance in PG with the genes we have examined. The role of the dopaminergic system in PG is highlighted by recent studies that showed robust association of development of PG and other impulsive behaviours following the use of DA agonists in those with Parkinson’s disease (reviewed by Voon & Fox, 2007). This led to collaboration with Dr. Antonio Strafella, in investigating the role of DA system genes in Parkinson’s disease patients who develop PG. This study is unique in the sense it will combine two powerful techniques (molecular genetics and functional imaging data) for the investigation of this phenomenon. Considering the results of our factor analysis, and the model incorporating DRD4 and the factors that may account for significant variance of PG (gambling problems, emotional vulnerability, impulsivity, cooperativeness/dependency, early wins, and erroneous beliefs), a few important remarks should be considered: 1. Emotional vulnerability refers to a large group of symptoms and psychiatric disorders, such as major depression, generalized anxiety disorder, and panic disorder. These disorders present significant genetic influence; therefore, we should keep in mind that the variance added to the model by this factor includes genetic factors that are not fully accounted for by DRD4. 2. Early wins and erroneous beliefs have been shown to be correlated with brain areas important in cognition and assigning value to rewards, such as the pre-frontal cortex and the limbic system; therefore, the next step would be to obtain objective measures of these cognitive functions (e.g., through neuropsychological testing) in order to uncover the genetic components. It has been reported, for example, that the COMT gene has an important effect on working memory in subjects diagnosed with other psychiatric and neurological conditions, such as Parkinson’s disease and schizophrenia. 3. Temperament factors, such as impulsivity, are known to be highly heritable; however, research has indicated that impulsivity is a complex concept, per se, that should be refined in order to better understand its relationship to psychiatric disorders. The next logical step is to refine the phenotype, examine the genetic component of related phenotypes, including addictive behaviours, such as nicotine and alcohol use, in relationship to genetic polymorphisms. PG is also highly comorbid with mood, anxiety, and personality disorders, and investigation into these co-occurring disorders (e.g., major depression

28 and antisocial personality disorder) may provide further refinement of the PG phenotype. As a future study, we intend to re-contact the study participants and administer questionnaires to collect additional measures and gather data, which will allow for sub-phenotype definition and genetic dissection. In summary, our results strongly suggest genetic vulnerability interacts with environmental factors, and psychiatric comorbidity is apparent in the development of PG. A more comprehensive model could be reached by testing for these interactions in a larger sample with detailed clinical assessment. We predict a more comprehensive model will provide important information that could be used for both prevention and treatment strategies.

Study Limitations

Sample Size One limitation of the study was the limited sample size. Genes vary across sub- populations, so to accurately determine the genetic and ethnic contributions to the disorder, we would benefit from a much larger sample. In particular, a larger sample of PGs with and without comorbid conditions (e.g., drug and alcohol dependence, major depression) is needed to better understand the genetic vulnerability in different groups of PGs (see Discussion section). We are in the process of collecting data from a larger sample.

Sample Selection Another source of limitation stems from the selection of the sample itself. None of the samples in this study were randomly drawn from the population. This means that we have a limited ability to generalize the results to the population; however a population-based study would be far more expensive than our clinical sample.

Sex Differences Previous reports have reported sex differences in the incidence of gambling behaviours (Blanco, Hasin, Petry, Stinson, & Grant, 2006). There could be different mechanisms influencing the etiology of PG in males and females. Biological differences, such as hormonal levels and patterns of brain development, may explain this observation. Another explanation may lie in disparate social roles and attitudes expected between the two genders.

Control Groups There were difficulties defining control groups/multiple control groups. The difficulty was that a study investigating the genetic factors of problem gambling requires multiple control groups. Since most people gamble in some form or another, the act of gambling in itself is not a disorder. The desire to gamble for fun, found in regular social gamblers, is likely linked to a sense of adventure and desire for excitement. Hence, social gamblers are not a disordered group. Nevertheless, social gambling is likely related to personality and temperament, and thus may be influenced by genetic factors. That being said, the traditional control group for genetic studies,

29 usually composed of unaffected or symptom-free subjects, may be systematically different from regular social gamblers. It is therefore important that in addition to a control group of non/infrequent gamblers, a control group of social gamblers be used in studies investigating the genetic factors of gambling. In fact, the first part of this research project, Special Problems in Gambling: Attention Deficit Hyperactivity Disorder (ADHD) and Pathways to Problem Gambling, was the first study to use healthy, non-problem, social gamblers as a control group. Exact gambling figures are difficult to obtain as they depend on how one defines gambling. It is estimated that approximately: 15% of the population report not gambling at all - group 0; 55% of the population gamble infrequently - group 1; 25% are regular social gamblers - group 2 (Note: what distinguishes a social versus infrequent gambler is hard to define precisely); 3.5% of the population sometimes gamble problematically - group 3 (Note: This group does not qualify for a clinical diagnosis of PG); 1.5% of the population qualifies for a clinical diagnosis of PG - group 4. The genetic makeup of all these groups is of interest, and groups 0, 1, and 2 would constitute as control groups. However, we considered the contrast of social gambler with problem gambler (2 vs. 4) and infrequent gambler with problem gambler (1 vs. 4) to be the most interesting. A current limitation is there was no definite cut off between the regular social gambler and infrequent gambler.

Measurement With all assessment tools, the chance of an incorrect assessment cannot be eliminated. We used the SOGS and CPGI (not used in the early samples), and a DSM-IV based questionnaire to determine gambling status. These measures have been found to be reliable and valid. We used these 3 instruments to reduce the chance of an incorrect assessment. In combining the measures, we had a high degree of certainty that the subjects were accurately placed into the pathological, problem and non-problem gambling (control) groups.

References

A.P.A. (1994). Diagnostic and Statistical Manual of Mental Disorders, 4th edition (DSM-IV) (4 ed.). Washington, DC: American Psychiatric Association. A.P.A. (2000). Diagnostic and Statistical Manual of Mental Disorders, 4th edition, text revision (DSM-IV-TR) Washington, DC: American Psychiatric Association. Accili, D., Fishburn, C. S., Drago, J., Steiner, H., Lachowicz, J. E., Park, B. H., et al. (1996). A targeted mutation of the D3 dopamine receptor gene is associated with hyperactivity in mice. Proc Natl Acad Sci U S A, 93(5), 1945-1949. Anney, R. J., Rees, M. I., Bryan, E., Spurlock, G., Williams, N., Norton, N., et al. (2002). Characterisation, mutation detection, and association analysis of alternative promoters and 5' UTRs of the human dopamine D3 receptor gene in schizophrenia. Mol Psychiatry, 7(5), 493-502. Asghari, V., Sanyal, S., Buchwaldt, S., Paterson, A., Jovanovic, V., & Van Tol, H. H. (1995). Modulation of intracellular cyclic AMP levels by different human dopamine D4 receptor variants. J Neurochem, 65(3), 1157-1165. Ballon, N., Leroy, S., Roy, C., Bourdel, M. C., Olie, J. P., Charles-Nicolas, A., et al. (2007). Polymorphisms TaqI A of the DRD2, BalI of the DRD3, exon III repeat of the DRD4, and 3' UTR VNTR of the DAT: association with childhood ADHD in male African- Caribbean cocaine dependents? Am J Med Genet B Neuropsychiatr Genet, 144B(8), 1034-1041. Barr, C. L., Wigg, K. G., Wu, J., Zai, C., Bloom, S., Tannock, R., et al. (2000). Linkage study of two polymorphisms at the dopamine D3 receptor gene and attention-deficit hyperactivity disorder. Am J Med Genet, 96(1), 114-117. Bellgrove, M. A., & Mattingley, J. B. (2008). Molecular genetics of attention. Ann N Y Acad Sci, 1129, 200-212. Benjamin, J., Li, L., Patterson, C., Greenberg, B. D., Murphy, D. L., & Hamer, D. H. (1996). Population and familial association between the D4 dopamine receptor gene and measures of Novelty Seeking. Nat Genet, 12(1), 81-84. Bergh, C., Eklund, T., Sodersten, P., & Nordin, C. (1997). Altered dopamine function in pathological gambling. Psychol Med, 27(2), 473-475. Beuten, J., Ma, J. Z., Payne, T. J., Dupont, R. T., Quezada, P., Huang, W., et al. (2005). Significant association of BDNF haplotypes in European-American male smokers but not in European-American female or African-American smokers. Am J Med Genet B Neuropsychiatr Genet, 139B(1), 73-80. Black, D. W., Monahan, P. O., Temkit, M., & Shaw, M. (2006). A family study of pathological gambling. Psychiatry Res, 141(3), 295-303. Black, D. W., Moyer, T., & Schlosser, S. (2003). Quality of life and family history in pathological gambling. J Nerv Ment Dis, 191(2), 124-126. Blanco, C., Hasin, D. S., Petry, N., Stinson, F. S., & Grant, B. F. (2006). Sex differences in subclinical and DSM-IV pathological gambling: results from the National Epidemiologic Survey on Alcohol and Related Conditions. Psychol Med, 36(7), 943-953. Blanco, C., Orensanz-Munoz, L., Blanco-Jerez, C., & Saiz-Ruiz, J. (1996). Pathological gambling and platelet MAO activity: a psychobiological study. Am J Psychiatry, 153(1), 119-121.

Blaszczynski, A., & Nower, L. (2002). A pathways model of problem and pathological gambling. Addiction, 97(5), 487-499. Blum, K., Sheridan, P. J., Wood, R. C., Braverman, E. R., Chen, T. J., & Comings, D. E. (1995). Dopamine D2 receptor gene variants: association and linkage studies in impulsive- addictive-compulsive behaviour. Pharmacogenetics, 5(3), 121-141. Bobb, A. J., Addington, A. M., Sidransky, E., Gornick, M. C., Lerch, J. P., Greenstein, D. K., et al. (2005). Support for association between ADHD and two candidate genes: NET1 and DRD1. Am J Med Genet B Neuropsychiatr Genet, 134(1), 67-72. Callegari-Jacques, S. M., Grattapaglia, D., Salzano, F. M., Salamoni, S. P., Crossetti, S. G., Ferreira, M. E., et al. (2003). Historical genetics: spatiotemporal analysis of the formation of the Brazilian population. Am J Hum Biol, 15(6), 824-834. Cavedini, P., Riboldi, G., Keller, R., D'Annucci, A., & Bellodi, L. (2002). Frontal lobe dysfunction in pathological gambling patients. Biol Psychiatry, 51(4), 334-341. Cetin, T., Freudenberg, F., Fuchtemeier, M., & Koch, M. (2004). Dopamine in the orbitofrontal cortex regulates operant responding under a progressive ratio of reinforcement in rats. Neurosci Lett, 370(2-3), 114-117. Chambers, R. A., & Potenza, M. N. (2003). Neurodevelopment, impulsivity, and adolescent gambling. J Gambl Stud, 19(1), 53-84. Chambers, R. A., Taylor, J. R., & Potenza, M. N. (2003). Developmental neurocircuitry of motivation in adolescence: a critical period of addiction vulnerability. Am J Psychiatry, 160(6), 1041-1052. Cloninger, C. R., Svrakic, D. M., & Przybeck, T. R. (1993). A psychobiological model of temperament and character. Arch Gen Psychiatry, 50(12), 975-990. Comings, D. E., & Blum, K. (2000). Reward deficiency syndrome: genetic aspects of behavioral disorders. Prog Brain Res, 126, 325-341. Comings, D. E., Gade-Andavolu, R., Gonzalez, N., Wu, S., Muhleman, D., Chen, C., et al. (2001). The additive effect of neurotransmitter genes in pathological gambling. Clin Genet, 60(2), 107-116. Comings, D. E., Gade, R., Wu, S., Chiu, C., Dietz, G., Muhleman, D., et al. (1997). Studies of the potential role of the dopamine D1 receptor gene in addictive behaviors. Mol Psychiatry, 2(1), 44-56. Comings, D. E., Gonzalez, N., Wu, S., Gade, R., Muhleman, D., Saucier, G., et al. (1999). Studies of the 48 bp repeat polymorphism of the DRD4 gene in impulsive, compulsive, addictive behaviors: Tourette syndrome, ADHD, pathological gambling, and substance abuse. Am J Med Genet, 88(4), 358-368. Comings, D. E., Gonzalez, N., Wu, S., Saucier, G., Johnson, P., Verde, R., et al. (1999). Homozygosity at the dopamine DRD3 receptor gene in cocaine dependence. Mol Psychiatry, 4(5), 484-487. Comings, D. E., Rosenthal, R. J., Lesieur, H. R., Rugle, L. J., Muhleman, D., Chiu, C., et al. (1996). A study of the dopamine D2 receptor gene in pathological gambling. Pharmacogenetics, 6(3), 223-234. da Silva Lobo, D. S., Vallada, H. P., Knight, J., Martins, S. S., Tavares, H., Gentil, V., et al. (2007). Dopamine genes and pathological gambling in discordant sib-pairs. J Gambl Stud, 23(4), 421-433. Daghestani, A. N., Elenz, E., & Crayton, J. W. (1996). Pathological gambling in hospitalized substance abusing veterans. J Clin Psychiatry, 57(8), 360-363.

Dalley, J. W., Laane, K., Theobald, D. E., Armstrong, H. C., Corlett, P. R., Chudasama, Y., et al. (2005). Time-limited modulation of appetitive Pavlovian memory by D1 and NMDA receptors in the nucleus accumbens. Proc Natl Acad Sci U S A, 102(17), 6189-6194. DeCaria, C. M., Hollander, E., Grossman, R., Wong, C. M., Mosovich, S. A., & Cherkasky, S. (1996). Diagnosis, neurobiology, and treatment of pathological gambling. J Clin Psychiatry, 57 Suppl 8, 80-83; discussion 83-84. Derevensky, J. L., & Gupta, R. (2000). Prevalence estimates of adolescent gambling: a comparison of the SOGS-RA, DSM-IV-J, and the GA 20 questions. J Gambl Stud, 16(2- 3), 227-251. Dick, D. M., Wang, J. C., Plunkett, J., Aliev, F., Hinrichs, A., Bertelsen, S., et al. (2007). Family-based association analyses of alcohol dependence phenotypes across DRD2 and neighboring gene ANKK1. Alcohol Clin Exp Res, 31(10), 1645-1653. Dudbridge, F. (2003). Pedigree disequilibrium tests for multilocus haplotypes. Genet Epidemiol, 25(2), 115-121. Ebstein, R. P., Novick, O., Umansky, R., Priel, B., Osher, Y., Blaine, D., et al. (1996). Dopamine D4 receptor (D4DR) exon III polymorphism associated with the human personality trait of Novelty Seeking. Nat Genet, 12(1), 78-80. Eisen, S. A., Lin, N., Lyons, M. J., Scherrer, J. F., Griffith, K., True, W. R., et al. (1998). Familial influences on gambling behavior: an analysis of 3359 twin pairs. Addiction, 93(9), 1375-1384. Elia, J., & Devoto, M. (2007). ADHD genetics: 2007 update. Curr Psychiatry Rep, 9(5), 434- 439. Folkman, S., & Lazarus, R. S. (1988a). Coping as a mediator of emotion. J Pers Soc Psychol, 54(3), 466-475. Folkman, S., & Lazarus, R. S. (1988b). The relationship between coping and emotion: implications for theory and research. Soc Sci Med, 26(3), 309-317. Friedel, S., Horro, F. F., Wermter, A. K., Geller, F., Dempfle, A., Reichwald, K., et al. (2005). Mutation screen of the brain derived neurotrophic factor gene (BDNF): identification of several genetic variants and association studies in patients with obesity, eating disorders, and attention-deficit/hyperactivity disorder. Am J Med Genet B Neuropsychiatr Genet, 132B(1), 96-99. Gambino, B. F., R.; Shaffer, H.; Renner, J.; Gourtnage, P. (1993). Perceived Family History of Problem Gambling and Scores on SOGS. Journal of Gambling Studies, 9(2), 169-184. Gauderman, W., & Morrison, J. (2006). QUANTO 1.1: A computer program for power and sample size calculations for genetic-epidemiology studies http://hydra.usc.edu/gxe. Gelernter, J., Yu, Y., Weiss, R., Brady, K., Panhuysen, C., Yang, B. Z., et al. (2006). Haplotype spanning TTC12 and ANKK1, flanked by the DRD2 and NCAM1 loci, is strongly associated to nicotine dependence in two distinct American populations. Hum Mol Genet, 15(24), 3498-3507. Guillin, O., Diaz, J., Carroll, P., Griffon, N., Schwartz, J. C., & Sokoloff, P. (2001). BDNF controls dopamine D3 receptor expression and triggers behavioural sensitization. Nature, 411(6833), 86-89. Gupta, R., & Derevensky, J. (1997). Familial and social influences on juvenile gambling behavior. J Gambl Stud, 13(3), 179-192.

Gupta, R., & Derevensky, J. L. (1998). Adolescent Gambling Behavior: A Prevalence Study and Examination of the Correlates Associated with Problem Gambling. J Gambl Stud, 14(4), 319-345. Habra, J., & Lee, G. (1995). Problem gambling and policy advice: the mutability and relative effects of structural, associational and attitudinal variables. Journal of Gambling Studies, 11, 105-121. Hall, D., Dhilla, A., Charalambous, A., Gogos, J. A., & Karayiorgou, M. (2003). Sequence variants of the brain-derived neurotrophic factor (BDNF) gene are strongly associated with obsessive-compulsive disorder. Am J Hum Genet, 73(2), 370-376. Heather, N., Tebbutt, J. S., Mattick, R. P., & Zamir, R. (1993). Development of a scale for measuring impaired control over alcohol consumption: a preliminary report. J Stud Alcohol, 54(6), 700-709. Heppner, P. P. (1988). The Problem Solving Inventory: Manual. Palo Alto, CA: Consulting Psychologists Press. Huang, W., Payne, T. J., Ma, J. Z., Beuten, J., Dupont, R. T., Inohara, N., et al. (2008). Significant Association of ANKK1 and Detection of a Functional Polymorphism with Nicotine Dependence in an African-American Sample. Neuropsychopharmacology. Ibanez, A., de Castro, I. P., Fernandez-Piqueras, J., Blanco, C., & Saiz-Ruiz, J. (2000). Pathological gambling and DNA polymorphic markers at MAO-A and MAO-B genes. Mol Psychiatry, 5(1), 105-109. Ibanez, A., Perez de Castro, I., Fernandez-Piqueras, J., & Saiz-Ruiz, J. (1999). Tyrosine hidroxilase gene and pathological gambling: an association study. Molecular Psychiatry, 4(suppl 1), S108-S109. Jacobs, D. F. (1989). Children of problem gamblers. Journal of Gambling Behavior, 5, 261-268. Jovanovic, V., Guan, H. C., & Van Tol, H. H. (1999). Comparative pharmacological and functional analysis of the human dopamine D4.2 and D4.10 receptor variants. Pharmacogenetics, 9(5), 561-568. Kelley, A. E. (2004). Ventral striatal control of appetitive motivation: role in ingestive behavior and reward-related learning. Neurosci Biobehav Rev, 27(8), 765-776. Kent, L., Green, E., Hawi, Z., Kirley, A., Dudbridge, F., Lowe, N., et al. (2005). Association of the paternally transmitted copy of common Valine allele of the Val66Met polymorphism of the brain-derived neurotrophic factor (BDNF) gene with susceptibility to ADHD. Mol Psychiatry, 10(10), 939-943. Kirley, A., Hawi, Z., Daly, G., McCarron, M., Mullins, C., Millar, N., et al. (2002). Dopaminergic system genes in ADHD: toward a biological hypothesis. Neuropsychopharmacology, 27(4), 607-619. Kopeckova, M., Paclt, I., Petrasek, J., Pacltova, D., Malikova, M., & Zagatova, V. (2008). Some ADHD polymorphisms (in genes DAT1, DRD2, DRD3, DBH, 5-HTT) in case-control study of 100 subjects 6-10 age. Neuro Endocrinol Lett, 29(2), 246-251. Kuntsi, J., McLoughlin, G., & Asherson, P. (2006). Attention deficit hyperactivity disorder. Neuromolecular Med, 8(4), 461-484. Le Foll, B., Diaz, J., & Sokoloff, P. (2003). Increased dopamine D3 receptor expression accompanying behavioral sensitization to nicotine in rats. Synapse, 47(3), 176-183. Le Foll, B., Diaz, J., & Sokoloff, P. (2005). A single cocaine exposure increases BDNF and D3 receptor expression: implications for drug-conditioning. Neuroreport, 16(2), 175-178.

Lee, M. S., & Ryu, S. H. (2002). No association between the dopamine D3 receptor gene and Korean alcohol dependence. Psychiatr Genet, 12(3), 173-176. Lesieur, H. R., & Blume, S. B. (1987). The South Oaks Gambling Screen (SOGS): a new instrument for the identification of pathological gamblers. Am J Psychiatry, 144(9), 1184- 1188. Lesieur, H. R., & Blume, S. B. (1990). Characteristics of pathological gamblers identified among patients on a psychiatric admissions service. Hosp Community Psychiatry, 41(9), 1009- 1012. Lesieur, H. R., Blume, S. B., & Zoppa, R. M. (1986). Alcoholism, drug abuse, and gambling. Alcohol Clin Exp Res, 10(1), 33-38. Lesieur, H. R., Cross, J., Frank, M., Welch, M., White, C. M., Rubenstein, G., et al. (1991). Gambling and pathological gambling among university students. Addict Behav, 16(6), 517-527. Lesieur, H. R., & Heineman, M. (1988). Pathological gambling among youthful multiple substance abusers in a therapeutic community. Br J Addict, 83(7), 765-771. Lesieur, H. R. R., J. (1989). Children of Gamblers Anonymous members. Journal of Gambling Behavior, 5, 269-281. Li, T., Liu, X., Zhao, J., Hu, X., Ball, D. M., Loh el, W., et al. (2002). Allelic association analysis of the dopamine D2, D3, 5-HT2A, and GABA(A)gamma2 receptors and serotonin transporter genes with heroin abuse in Chinese subjects. Am J Med Genet, 114(3), 329-335. Limosin, F., Romo, L., Batel, P., Ades, J., Boni, C., & Gorwood, P. (2005). Association between dopamine receptor D3 gene BalI polymorphism and cognitive impulsiveness in alcohol- dependent men. Eur Psychiatry, 20(3), 304-306. Littman-Sharp, N., & Stirpe, T. (1997). The Inventory of Gambling Situations: An interim report on a risk assessment instrument. Paper presented at the 2nd Bi-Annual Ontario Conference on Problem and Compulsive Gambling, Toronto, Ontario. Liu, Q. R., Walther, D., Drgon, T., Polesskaya, O., Lesnick, T. G., Strain, K. J., et al. (2005). Human brain derived neurotrophic factor (BDNF) genes, splicing patterns, and assessments of associations with substance abuse and Parkinson's Disease. Am J Med Genet B Neuropsychiatr Genet, 134B(1), 93-103. Misener, V. L., Luca, P., Azeke, O., Crosbie, J., Waldman, I., Tannock, R., et al. (2004). Linkage of the dopamine receptor D1 gene to attention-deficit/hyperactivity disorder. Mol Psychiatry, 9(5), 500-509. Muglia, P., Jain, U., & Kennedy, J. L. (2002). A transmission disequilibrium test of the Ser9/Gly dopamine D3 receptor gene polymorphism in adult attention-deficit hyperactivity disorder. Behav Brain Res, 130(1-2), 91-95. Murray, J. B. (1993). Review of research on pathological gambling. Psychol Rep, 72(3 Pt 1), 791-810. N.O.R.C. (1999). National Opinion Research Center - Gambling Impact and Behavior Study, Report to the National Gambling Impact Study Commission. Nazarian, A., Russo, S. J., Festa, E. D., Kraish, M., & Quinones-Jenab, V. (2004). The role of D1 and D2 receptors in the cocaine conditioned place preference of male and female rats. Brain Res Bull, 63(4), 295-299.

Neville, M. J., Johnstone, E. C., & Walton, R. T. (2004). Identification and characterization of ANKK1: a novel kinase gene closely linked to DRD2 on chromosome band 11q23.1. Hum Mutat, 23(6), 540-545. Nicodemus, K. K., Liu, W., Chase, G. A., Tsai, Y. Y., & Fallin, M. D. (2005). Comparison of type I error for multiple test corrections in large single-nucleotide polymorphism studies using principal components versus haplotype blocking algorithms. BMC Genet, 6 Suppl 1, S78. Ohtsuka, K., Bruton, E., DeLuca, L., & Borg, V. (1997). Sex differences in pathological gambling using gaming machines. Psychol Rep, 80(3 Pt 1), 1051-1057. Patton, J. H., Stanford, M. S., & Barratt, E. S. (1995). Factor structure of the Barratt impulsiveness scale. J Clin Psychol, 51(6), 768-774. Payton, A., Holmes, J., Barrett, J. H., Hever, T., Fitzpatrick, H., Trumper, A. L., et al. (2001). Examining for association between candidate gene polymorphisms in the dopamine pathway and attention-deficit hyperactivity disorder: a family-based study. Am J Med Genet, 105(5), 464-470. Perez de Castro, I., Ibanez, A., Saiz-Ruiz, J., & Fernandez-Piqueras, J. (1999). Genetic contribution to pathological gambling: possible association between a functional DNA polymorphism at the serotonin transporter gene (5-HTT) and affected men. Pharmacogenetics, 9(3), 397-400. Perez de Castro, I., Ibanez, A., Saiz-Ruiz, J., & Fernandez-Piqueras, J. (2002). Concurrent positive association between pathological gambling and functional DNA polymorphisms at the MAO-A and the 5-HT transporter genes. Mol Psychiatry, 7(9), 927-928. Perez de Castro, I., Ibanez, A., Torres, P., Saiz-Ruiz, J., & Fernandez-Piqueras, J. (1997). Genetic association study between pathological gambling and a functional DNA polymorphism at the D4 receptor gene. Pharmacogenetics, 7(5), 345-348. Pilla, M., Perachon, S., Sautel, F., Garrido, F., Mann, A., Wermuth, C. G., et al. (1999). Selective inhibition of cocaine-seeking behaviour by a partial dopamine D3 receptor agonist. Nature, 400(6742), 371-375. Potenza, M. N. (2001). The neurobiology of pathological gambling. Semin Clin Neuropsychiatry, 6(3), 217-226. Potenza, M. N., Xian, H., Shah, K., Scherrer, J. F., & Eisen, S. A. (2005). Shared genetic contributions to pathological gambling and major depression in men. Arch Gen Psychiatry, 62(9), 1015-1021. Potter, D. M. (2006). Omnibus permutation tests of the association of an ensemble of genetic markers with disease in case-control studies. Genet Epidemiol, 30(5), 438-446. Retz, W., Rosler, M., Supprian, T., Retz-Junginger, P., & Thome, J. (2003). Dopamine D3 receptor gene polymorphism and violent behavior: relation to impulsiveness and ADHD- related psychopathology. J Neural Transm, 110(5), 561-572. Reuter, J., Raedler, T., Rose, M., Hand, I., Glascher, J., & Buchel, C. (2005). Pathological gambling is linked to reduced activation of the mesolimbic reward system. Nat Neurosci, 8(2), 147-148. Risch, N. (1990). Linkage strategies for genetically complex traits. I. Multilocus models. Am J Hum Genet, 46(2), 222-228. Roy, A., Adinoff, B., Roehrich, L., Lamparski, D., Custer, R., Lorenz, V., et al. (1988). Pathological gambling. A psychobiological study. Arch Gen Psychiatry, 45(4), 369-373.

Sanchez, C. J., Bailie, T. M., Wu, W. R., Li, N., & Sorg, B. A. (2003). Manipulation of dopamine d1-like receptor activation in the rat medial prefrontal cortex alters stress- and cocaine-induced reinstatement of conditioned place preference behavior. Neuroscience, 119(2), 497-505. Sander, T., Harms, H., Podschus, J., Finckh, U., Nickel, B., Rolfs, A., et al. (1995). Dopamine D1, D2 and D3 receptor genes in alcohol dependence. Psychiatr Genet, 5(4), 171-176. Scherrer, J. F., Xian, H., Shah, K. R., Volberg, R., Slutske, W., & Eisen, S. A. (2005). Effect of genes, environment, and lifetime co-occurring disorders on health-related quality of life in problem and pathological gamblers. Arch Gen Psychiatry, 62(6), 677-683. Self, D. W. (2004). Regulation of drug-taking and -seeking behaviors by neuroadaptations in the mesolimbic dopamine system. Neuropharmacology, 47 Suppl 1, 242-255. Shaffer, H. J., & Hall, M. N. (2001). Updating and refining prevalence estimates of disordered gambling behaviour in the United States and Canada. Can J Public Health, 92(3), 168- 172. Shaffer, H. J., Hall, M. N., & Vander Bilt, J. (1999). Estimating the prevalence of disordered gambling behavior in the United States and Canada: a research synthesis. Am J Public Health, 89(9), 1369-1376. Sivagnanasundaram, S., Morris, A. G., Gaitonde, E. J., McKenna, P. J., Mollon, J. D., & Hunt, D. M. (2000). A cluster of single nucleotide polymorphisms in the 5'-leader of the human dopamine D3 receptor gene (DRD3) and its relationship to schizophrenia. Neurosci Lett, 279(1), 13-16. Slutske, W. S., Eisen, S., True, W. R., Lyons, M. J., Goldberg, J., & Tsuang, M. (2000). Common genetic vulnerability for pathological gambling and alcohol dependence in men. Arch Gen Psychiatry, 57(7), 666-673. Slutske, W. S., Eisen, S., Xian, H., True, W. R., Lyons, M. J., Goldberg, J., et al. (2001). A twin study of the association between pathological gambling and antisocial personality disorder. J Abnorm Psychol, 110(2), 297-308. Smith, G. J., & Wynne, H. J. (2002). Measuring gambling and problem gambling in Alberta using the Canadian Problem Gambling Index (CPGI): Final report. . Edmonton, AB: Alberta Gaming Research Institute. Sokoloff, P., Giros, B., Martres, M. P., Bouthenet, M. L., & Schwartz, J. C. (1990). Molecular cloning and characterization of a novel dopamine receptor (D3) as a target for neuroleptics. Nature, 347(6289), 146-151. Staddon, S., Arranz, M. J., Mancama, D., Perez-Nievas, F., Arrizabalaga, I., Anney, R., et al. (2005). Association between dopamine D3 receptor gene polymorphisms and schizophrenia in an isolate population. Schizophr Res, 73(1), 49-54. Strobel, A., Wehr, A., Michel, A., & Brocke, B. (1999). Association between the dopamine D4 receptor (DRD4) exon III polymorphism and measures of Novelty Seeking in a German population. Mol Psychiatry, 4(4), 378-384. Szczepankiewicz, A., Dmitrzak-Weglarz, M., Skibinska, M., Slopien, A., Leszczynska- Rodziewicz, A., Czerski, P., et al. (2007). Study of dopamine receptors genes polymorphisms in bipolar patients with comorbid alcohol abuse. Alcohol Alcohol, 42(2), 70-74. Tanabe, J., Thompson, L., Claus, E., Dalwani, M., Hutchison, K., & Banich, M. T. (2007). Prefrontal cortex activity is reduced in gambling and nongambling substance users during decision-making. Hum Brain Mapp, 28(12), 1276-1286.

Toneatto, T. (1999). Cognitive psychopathology of problem gambling. Subst Use Misuse, 34(11), 1593-1604. Turner, N. (2006). How many paths are there? Use of structural equation modelling to examine risk factors for pathological gambling. Paper presented at the 13th International Conference on Gambling and Risk Taking, Lake Tahoe, NV. Uhl, G. R., Liu, Q. R., Walther, D., Hess, J., & Naiman, D. (2001). Polysubstance abuse- vulnerability genes: genome scans for association, using 1,004 subjects and 1,494 single- nucleotide polymorphisms. Am J Hum Genet, 69(6), 1290-1300. Vachon, J., Vitaro, F., Wanner, B., & Tremblay, R. E. (2004). Adolescent gambling: relationships with parent gambling and parenting practices. Psychol Addict Behav, 18(4), 398-401. Van Craenenbroeck, K., Clark, S. D., Cox, M. J., Oak, J. N., Liu, F., & Van Tol, H. H. (2005). Folding efficiency is rate-limiting in dopamine D4 receptor biogenesis. J Biol Chem, 280(19), 19350-19357. Van Tol, H. H., Wu, C. M., Guan, H. C., Ohara, K., Bunzow, J. R., Civelli, O., et al. (1992). Multiple dopamine D4 receptor variants in the human population. Nature, 358(6382), 149-152. Volberg, R. A., & Steadman, H. J. (1988). Refining prevalence estimates of pathological gambling. Am J Psychiatry, 145(4), 502-505. Voon, V., & Fox, S. H. (2007). Medication-related impulse control and repetitive behaviors in Parkinson disease. Arch Neurol, 64(8), 1089-1096. Walters, G. D., & Contri, D. (1998). Outcome Expectancies for Gambling: Empirical Modeling of a Memory Network in Federal Prison Inmates. J Gambl Stud, 14(2), 173-191. Wiesbeck, G. A., Dursteler-MacFarland, K. M., Wurst, F. M., Walter, M., Petitjean, S., Muller, S., et al. (2006). No association of dopamine receptor sensitivity in vivo with genetic predisposition for alcoholism and DRD2/DRD3 gene polymorphisms in alcohol dependence. Addict Biol, 11(1), 72-75. Winters, K. C., & Rich, T. (1999). A twin study of adult gambling behavior. Journal of Gambling Studies, 14(3), 213-225. Yang, B. Z., Kranzler, H. R., Zhao, H., Gruen, J. R., Luo, X., & Gelernter, J. (2007). Association of haplotypic variants in DRD2, ANKK1, TTC12 and NCAM1 to alcohol dependence in independent case control and family samples. Hum Mol Genet, 16(23), 2844-2853. Zack, M., & Poulos, C. X. (2004). Amphetamine primes motivation to gamble and gambling- related semantic networks in problem gamblers. Neuropsychopharmacology, 29(1), 195- 207. Zuckerman, M. (1988). Sensation seeking and behavior disorders. Arch Gen Psychiatry, 45(5), 502-504. Zung, W. W. (1965). A Self-Rating Depression Scale. Arch Gen Psychiatry, 12, 63-70.