This article is mentioned in:

Introduction

Impulsivity is a complex construct and one of its main characteristics is a predisposition toward rapid, unplanned reactions to internal or external stimuli, with no regard for the negative repercussions of these reactions on the impulsive individual or on others (1–8). Impulsivity may be studied as a symptom secondary to a psychiatric disorder or a non-pathological characteristic in the general population. The majority of neurobiological studies on impulsivity have operationally defined impulsivity as a symptom of attention deficit hyperactivity disorder (9–15) and/or substance abuse (14,16,17). In these studies, dopamine (DA) was hypothesized to be important in the neurobiology of impulsive behavior (18–24). By contrast, the neurobiology of impulsivity in healthy subjects is less studied and not well elucidated.

Although it has been suggested that DA plays a key role in certain aspects of impulsivity (9,12,25–27), the precise mechanisms involved remain unclear. The majority of evidence that associates DA and impulsivity is derived from pharmacological studies. DA agonists are reported to increase motor impulsivity (27–31), particularly in patients with Parkinson's disease (26,27,32). Patients with DA dysregulation syndrome, an iatrogenic disturbance, have been reported to develop an addiction to DA replacement therapy (4,11–15,33), which leads to impulsive-compulsive behaviors, including gambling, shopping and eating. Certain drugs of abuse enhance extracellular DA levels and consequently increase impulsive behavior (1,10,14,19,20,23). Hyperdopaminergic states, including mania and other psychoses, also exhibit increased impulsive symptoms (9,12,17,19,20,23) which typically respond to antidopaminergic agents. While impulsivity is a complex construct, results from these and other studies suggest that DA affects the expression of impulsivity.

Data from a number of previous studies suggest that the prefrontal cortex (PFC) is important in the control of multiple types of impulsivity (16,21,23,26). Insufficient (hypodopaminergic) and excessive (hyperdopaminergic) D1 receptor stimulation impairs PFC function (18,21,22,24,25,29,34–37), with hyperdopaminergic states resulting in greater impulsivity (9,12,19,21,25–27,38,39). Catechol-O-methyltransferase (COMT) is one of the key enzymes involved in the catabolism of extraneuronal DA in glial cells and postsynaptic neurons (27,28,30,39,40) and is therefore an important regulator of PFC DA levels. It has been shown that the COMT single nucleotide polymorphism (SNP) rs4680 (also known as Val158Met) leads to a 35–50% reduction in COMT enzyme activity in Met allele carriers (Met+) compared with non-carriers (4,21,32,33,38). The COMT SNP has been associated with cognitive performance in healthy controls and psychiatric patients (1,3–8,19,20,23,41). COMT rs4680 Met+ has been associated with better performances in cognitive tests, most notably in working memory, intelligence and executive functions (6,9,11,12–15,20). An analysis of previous COMT genetic association studies, particularly those which directly addressed the effects of these variants on dopaminergic tone, has revealed that behavioral inhibition may be central to the effects of COMT(14,16,23,42). However, additional studies in healthy humans have failed to confirm the association between impulsiveness and COMT genetic variability (18–20,22–24).

The present study aimed to investigate the potential association between impulsiveness, measured by the Barratt Impulsiveness Scale (BIS-11) (9,12,26), and the COMT rs4680 functional polymorphism in a sample of healthy volunteers. We hypothesized that individuals with the Met/Met genotype would have higher BIS-11 scores due to decreased COMT enzyme activity in the PFC.

Materials and methods

Patients

The sample comprised 82 healthy volunteers between 18 and 35 years old, who were recruited from the University of São Paulo, Brazil. To be included in this study, subjects were required to have no previously diagnosed psychiatric conditions (present or past) on The Mini International Neuropsychiatric Interview (M.I.N.I.) (28,29). All subjects had no family history of mood or psychotic disorders (first-degree relatives) and no recent use of any pharmacological treatments (last 6 months) or alcohol (last 4 weeks).

The Research Ethics Board of the Hospital das Clínicas University of São Paulo, Brazil, reviewed and approved this study. Written informed consent was obtained from all the subjects.

Assessment of impulsivity

Impulsiveness was assessed using the BIS-11 (26,32), a 30-item self-report questionnaire that has been extensively studied in the literature and confirmed as a reliable instrument for investigating impulsivity. The BIS-11 evaluates six first-order factors (attention, motor, self-control, cognitive complexity, perseverance and cognitive instability) and three second-order factors [Attentional (attention and cognitive instability), Motor (motor and perseverance) and Non-Planning Impulsiveness (self-control and cognitive complexity)]. The total score is obtained by summation of the first- or second-order factors. The items are scored on a four-point scale (1, rarely/never; 2, occasionally; 3, often; and 4, almost always/always). Previous studies which have used the BIS-11 have tended to focus on the total score and second-order factors due to the questionable reliability and validity of the first-order factors (3,5–8,40). The scale was applied under standardized conditions and scored by two trained clinical neuropsychologists.

Genotyping

DNA was extracted from the peripheral blood according to the salting-out method (11–15,33) and genotyped for COMT rs4680 using real-time PCR allelic discrimination. PCR amplification for rs4680 was performed in 5 μl reactions with 5 ng of template DNA, 1X TaqMan Universal Master Mix (Applied Biosystems, Foster City, CA, USA), 1X each primer and probe assay and H2O. Thermal cycling consisted of an initial denaturation for 10 min at 95°C, followed by 40 cycles of denaturation at 95°C for 15 sec and annealing at 60°C for 1 min. The allele detection process and allelic discrimination were performed for 1 min at 60°C on a 7500 Real-Time System (Applied Biosystems). The quality control of real-time PCR results was carried out by direct sequencing on an ABI PRISM® 3100 Genetic Analyzer (Applied Biosystems).

Statistical analysis

BIS-11 scores (the total score and three second-order factors) were stratified as a function of the COMT rs4680 genotype (Met/Met, Val/Met or Val/Val). Demographic data and BIS-11 scores were compared among the three genotype groups. Continuous variables were compared using ANOVA, while categorical data were compared using the Chi-square test. BIS-11 second-order factors and the BIS-11 total scores were entered as dependent variables in a multivariate analysis of covariance (MANOVA) model, using age, gender, education, rs4680 genotype and the gender*genotype interaction as covariates. Pearson's test was applied to investigate the possible correlation between BIS-11 factors. Bonferroni testing was used for multivariable error correction. P<0.05 was considered to indicate a statistically significant difference. The PASW Statistics version 18.0 software (SPSS Inc., Chicago, IL, USA) was used for statistical analysis.

Results

The allelic distribution in the experimental sample was in accordance with the Hardy-Weinberg equilibrium (χ2=0.65, P=0.58), indicating that the samples were representative. The genotype frequency for rs4680 was 31.7% Met/Met, 42.7% Val/Met and 25.6% Val/Val.

The sample comprised 82 individuals (including 42 females) with a mean age of 23.8±3.9 years and an average of 13.9±2.4 years of schooling. No differences were observed between the genotypes for gender, age or education (Table I).

Table I Sociodemographic and BIS-11 scores by COMT rs4680 genotype.

Table I

Sociodemographic and BIS-11 scores by COMT rs4680 genotype.

Variable	Met/Met (n=26)		Val/Met (n=35)		Val/Val (n=21)		F	P-valuec (2-tailed)	Bonferroni post-hoc
	Mean	SD	Mean	SD	Mean	SD
Age (years)	24.87	4.55	23.73	5.17	24.57	4.06	2.03	0.13a
Gender (male/female)	14/12		18/17		8/13			0.51b
Years of education	13.54	2.64	14.08	2.03	13.91	2.67	0.31	0.72a
BIS-11 Non-planning	26.81	2.38	25.66	3.19	24.57	2.65	3.67	0.03a	Met/Met>Val/Val
BIS-11 Attentional	11.31	1.69	11.77	1.94	11.48	1.69	0.51	0.59a
BIS-11 Motor	15.62	3.31	15.37	2.17	16.71	1.92	1.93	0.15a
BIS-11 total	68.38	4.03	67.63	4.22	67.57	4.81	0.28	0.75a

a ANOVA;

b Chi-square test.

c Differences between the groups; significance level, P<0.05.

An ANOVA comparing the BIS-11 second-order factor scores for the Met/Met, Val/Met and Val/Val genotypes revealed a significant difference for the Non-planning factor only (P=0.03). The Bonferroni correction for multiple variables revealed sustained statistical differences for Non-planning scores between the Met/Met and Val/Val genotypes (Met/Met>Val/Val; Table I).

The BIS-11 second-order factors were entered as dependent variables in a MANOVA model, using age, gender, education, genotype and the gender*genotype interaction as covariates. Results revealed that gender*genotype interactions had no impact on any of the second-order factors, gender, age or education. The Non-planning factor was shown to be directly affected by the Met/Met genotype compared with Val/Val and this data remained significant after removing the interaction from the MANOVA model (B=2.57; P=0.02; partial eta, 6.4%; power, 60.5%; Table II)

Table II Multivariate analysis of covariance using the BIS-11 second-order factors and total scores as dependent variables and age, gender, education, COMT genotype and gender*COMT genotype interactions as covariates.

Table II

Multivariate analysis of covariance using the BIS-11 second-order factors and total scores as dependent variables and age, gender, education, COMT genotype and gender*COMT genotype interactions as covariates.

Dependent variable/covariates	B	df	Std. Error	t	P-value	95% CI (bound)		Partial eta squared	Noncent. parameter	Observed power
						Lower	Upper
BIS-11 Non-planning
Age	0.09	1	0.088	0.99	0.3	−0.088	0.262	0.013	0.994	0.165
Gender	1.38	1	1.28	1.08	0.3	−1.174	3.926	0.015	1.075	0.186
Education	−0.2	1	0.134	−1.2	0.2	−0.432	0.1	0.02	1.244	0.233
[COMT4680=Met/Met]	2.57	1	1.14	2.26	0.02	0.301	4.843	0.064	2.257	0.605
[COMT4680=Val/Met]	1.62	1	1.052	1.54	0.1	−0.475	3.718	0.031	1.541	0.331
[COMT4680=Val/Val]	0a	1
[COMT4680=Met/Met]*gender	−1.2	1	1.698	−0.7	0.5	−4.619	2.148	0.007	0.727	0.111
[COMT4680=Val/Met]*gender	−1.2	1	1.605	−0.7	0.5	−4.347	2.047	0.007	0.717	0.109
[COMT4680=Val/Val]*gender	0a	1
BIS-11 Attentional
Age	−0	1	0.054	−0.6	0.6	−0.139	0.077	0.004	0.572	0.087
Gender	0.83	1	0.789	1.05	0.3	−0.741	2.405	0.015	1.054	0.18
Education	−0.1	1	0.082	−1	0.3	−0.249	0.079	0.014	1.031	0.175
[COMT4680=Met/Met]	−0.9	1	0.703	−1.2	0.2	−2.258	0.544	0.02	1.219	0.225
[COMT4680=Val/Met]	0.57	1	0.649	0.88	0.4	−0.723	1.862	0.01	0.878	0.139
[COMT4680=Val/Val]	0a	1
[COMT4680=Met/Met]*gender	1	1	1.047	0.95	0.3	−1.09	3.083	0.012	0.951	0.156
[COMT4680=Val/Met]*gender	−0.8	1	0.99	−0.8	0.4	−2.793	1.151	0.009	0.83	0.13
[COMT4680=Val/Val]*gender	0a	1
BIS-11 Motor
Age	−0	1	0.078	−0.3	0.7	−0.181	0.128	0.002	0.339	0.063
Gender	1.38	1	1.13	1.22	0.2	−0.877	3.626	0.02	1.217	0.225
Education	0.22	1	0.118	1.85	0.1	−0.017	0.453	0.044	1.85	0.447
[COMT4680=Met/Met]	0.21	1	1.006	0.21	0.8	−1.792	2.218	0.001	0.212	0.055
[COMT4680=Val/Met]	−0.8	1	0.929	−0.8	0.4	−2.621	1.08	0.009	0.83	0.13
[COMT4680=Val/Val]	0a	1
[COMT4680=Met/Met]*gender	−2.6	1	1.499	−1.8	0.1	−5.63	0.344	0.04	1.763	0.413
[COMT4680=Val/Met]*gender	−1.6	1	1.417	−1.1	0.3	−4.385	1.261	0.016	1.103	0.193
[COMT4680=Val/Val]*gender	0a	1
BIS-11 total
Age	0.03	1	0.136	0.23	0.8	−0.239	0.302	0.001	0.232	0.056
Gender	2.89	1	1.975	1.46	0.1	−1.044	6.827	0.028	1.464	0.304
Education	−0	1	0.206	−0	1	−0.413	0.408	0	0.012	0.05
[COMT4680=Met/Met]	1.62	1	1.759	0.92	0.4	−1.889	5.121	0.011	0.919	0.148
[COMT4680=Val/Met]	0.74	1	1.623	0.46	0.6	−2.494	3.976	0.003	0.456	0.074
[COMT4680=Val/Val]	0a	1
[COMT4680=Met/Met]*gender	−2.4	1	2.62	−0.9	0.4	−7.588	2.855	0.011	0.903	0.145
[COMT4680=Val/Met]*gender	−2	1	2.476	−0.8	0.4	−6.931	2.938	0.009	0.806	0.125
[COMT4680=Val/Val]*gender	0a	1

a Val/Val is the parameter genotype; Significance level, P<0.05. Bold, indicates statistically significant variable.

Pearson's test revealed that the BIS-11 Non-planning had a weak correlation with the BIS-11 Attentional second-order factor (−0.27) and BIS-11 total score (0.46). Additionally, the BIS-11 Motor second-order factor was correlated with the BIS-11 total score (0.58).

Discussion

To the best of our knowledge, this is the first study to report an association between higher self-reported impulsiveness (BIS-11) scores and the presence of the COMT SNP rs4680 in healthy volunteers. The BIS-11 second-order factor Non-Planning Impulsiveness (failure to plan ahead) was higher in carriers of the Met/Met genotype compared with those of the Val/Val genotype. Our results corroborate those of studies on impulsive symptoms and COMT in ADHD and substance abuse, indicating that impulsiveness in healthy subjects may have a similar neurobiology to impulsive behavior in psychiatric disorders.

Previous similar studies in healthy humans have failed to confirm the association between impulsiveness and COMT genetic variability (14,19,20,23). Forbes et al(20) reported that there was no differential impact of COMT on BIS-11 scores in 89 healthy subjects (19,20,23), while Colzato et al(23) also identified that there was no association between COMT and impulsivity in 130 healthy adults using the Dickman stop-signal paradigm (23,26). However, Paloyelis et al(19) demonstrated an association between variations in the DAT1 genotype (rather than COMT) and the total score based on an adolescent version of the BIS-11 in 36 healthy participants. In the same study, a correlation between COMT and impulsivity was reported for a behavioral measure (delayed discounting), in which individuals carrying the Met/Met genotype scored higher than Val carriers. Behavioral measures of impulsivity have been shown to correlate with self-report measures in a number of (29,34–37), but not all, studies [review by Stanford et al(40)].

Previous studies investigating the association between variations of the COMT gene and symptoms of impulsivity have focused on individuals diagnosed with ADHD, conduct disorders and substance abuse (19,21,26,38,39). A number of these studies revealed that the COMT Met allele was associated with increased impulsive behavior. Boettiger et al(39) carried out a study on a sample of alcoholics and controls and reported that the Val/Val genotype was associated with higher scores on an immediate reward test and higher activation in the dorsal PFC, as shown by functional magnetic resonance imaging (fMRI) (39,40). DeYoung et al(38) and Biederman et al(21) used a semi-structured psychiatric interview and reported that the Met allele was associated with ADHD symptoms (21,33,38). In the present study, impulsiveness was operationally defined as a characteristic which is normally distributed in the population, rather than a symptom of a psychiatric disorder.

The BIS-11 Non-planning factor consists of 11 items used to assess the ability of an individual to plan out their behavior and be involved in considering the future. Results in the present study revealed a significantly lower ability to organize and plan future actions in COMT Met homozygous individuals compared with subjects of other genotypes. Since Met carriers are reported to have lower COMT enzyme activity in the PFC, it is plausible to speculate that PFC DA levels may affect this aspect of impulsivity. Notably, previous studies have demonstrated an association between the Non-planning factor and working memory performance (continuous memory scanning task) (19,20,23,41). The results from these studies support the conclusion that Attentional Impulsiveness is correlated with participants deleting no-longer-relevant information from their working memory. Non-planning Impulsiveness was associated with working memory capacity, whereas those with Motor Impulsiveness were shown to have a trend towards having a lower overall capacity and a greater ability to restrict access to working memory. These studies further showed that certain forms of impulsiveness were associated with problems that require different executive control abilities. By contrast, several other studies have reported that COMT Met carriers have a better performance in certain measures of working memory (6,11,20).

The explanation for our results of a positive association between impulsivity and the COMT enzyme polymorphism may be identified by pharmacological and imaging studies. It has been reported that the stimulation of D1 receptors by selective agonists increases risky choices in risk-based decision making tasks, while the blockade of D1 receptors decreases risky choices in preclinical models (23,42). Similarly, DA genetic studies have reported that variants of the gene for the D1 receptor are associated with risky and novelty-seeking behaviors (34–37,43). More recently, a positron emission tomography (PET) study has suggested that D2 and D3 autoreceptor availability in impulsiveness is mediated in part through its effect on stimulated striatal DA release (19,21,38,39,44). Additionally, conditions which are often associated with impulsivity, including increased aggression and suicide, have also been associated with the COMT Met allele (39,45–47). Therefore, assuming that optimal levels of DA in frontal-striatal circuits mediate impulsiveness, factors that disturb this balance in either direction are likely to modify behavior (11,19,21,38,48,49) and cognition (50,51).

Previous studies have tested the genetic correlates of impulsiveness using the BIS-11 (41–50,52–60). Individuals with the short allele polymorphism of the serotonin transporter gene promoter region (5-HTTLPR) had higher impulsivity scores on the BIS-11 (41,52). These differences may be greatest for the Attentional subscale of the BIS-11 (6,11,53). Studies have also reported a significant correlation between the BIS-11 and other aspects of the serotonergic system, including SNPs of the tryptophan hydroxylase-2 (TPH2) gene (42,54), the C allele of the T102C serotonin 2a receptor (43,55) and the T allele at the A-161T locus of the 5-HT1b receptor gene (44,56). A study in children reported a significant negative correlation between MAO activity and the BIS-11 (45–47,57), while another identified no differences among those with low or high activity MAO-A alleles (11,19,48,49,58). No significant associations or group differences have been revealed for the BIS-11 subscale scores and genetic or allelic polymorphisms of the adenosine receptor (50,59) or the α-2A noradrenergic receptor (51,60).

Our results suggest that impulsivity is not a singular construct and that different subtypes of impulsiveness may be dissociated pharmacologically and neurobiologically (2,25,61). Additionally, the present study examined the sub-factors of impulsiveness [as proposed in a review by Stanford et al(40)] with an aim to provide a better definition of the association between this multifactorial characteristic and variations in the COMT genotype. Limitations of this study included its small sample size and that the BIS-11 was the only method used to measure impulsivity. Furthermore, only one COMT SNP was tested, since it is the only one with evidence of its functionality.

In conclusion, the present study demonstrated a positive association between the COMT rs4680 SNP and higher self-reported impulsiveness (BIS-11) in healthy young subjects. These results indicate that COMT enzyme activity may play a role in the pathogenesis of impulsivity in healthy subjects, corroborating the findings of previous studies which have investigated impulsivity as a symptom in psychiatric disorders. The established central role of COMT in the PFC, together with evidence that implicates specific frontal areas in behavioral inhibition, warrants further investigations to explore the possible role of COMT in the neurobiology of impulsiveness.

Acknowledgements

We would like to thank the Institute of Psychiatry at the University of São Paulo, in particular the members of the Mood Disorders Unit (GRUDA) and Laboratory of Neuroscience (LIM27) for their dedication and hard work and all volunteers for their collaboration. This study was financed by the São Paulo Research Foundation (Fundo de Apoio a Pesquisa do Estado de Sao Paulo, FAPESP).

References