Dopaminergic genes modulate response inhibition in alcohol abusing adults

Abstract

Compulsion in alcohol use disorders (AUD) has been attributed to impairment in response inhibition. Because genes that regulate dopamine (DA) have been implicated not only for risk for AUD but also for impulsivity based on behavioral studies, we set out to examine the underlying neural mechanisms associated with these effects. We collected functional magnetic resonance imaging images on 53 heavy drinking but otherwise healthy adults while performing the Go/NoGo task. We predicted that genetic variants previously reported in the literature to be associated with substance abuse, specifically the DRD2 rs1799732 and DRD4 VNTR, will modulate neural processes underlying response inhibition. Our results showed differential neural response for the DRD4 VNTR during successful inhibition in the inferior frontal gyrus (IFG) (cluster-corrected P < 0.05, z = 1.9). Similarly, DRD2 rs1799732 groups were significantly different in the precuneus and cingulate gyrus during successful response inhibition (cluster-corrected P < 0.05, z = 1.9). These findings provide further evidence for the role of DAergic genes in modulating neural response in areas that underlie response inhibition and self-monitoring processes. Variants within these genes appear to influence processes related to impulsive behavior, which may increase one’s risk for alcohol abuse and dependence.

INTRODUCTION

Over the past two decades, there has been considerable interest in the question of whether dopamine (DA) genes play an important role in alcohol abuse and dependence, and inhibitory control. A recent study reported positive associations between dopamine genes and measures of alcohol dependence as well as inhibitory control (Hack et al. 2010). Historically, variations in two genes, the DRD2 and DRD4, have been examined more than any other gene. Evidence for the importance of these two genes has been mixed when examining clinical phenotypes. However, studies using endophenotypes related to craving or inhibitory control have suggested an association. For example, the DRD4 variable number of tandem repeat polymorphism (DRD4 VNTR) has been shown to modulate processes related to alcohol craving and reward. In our previous work, we found an association between DRD4 VNTR and the neural mechanisms in response to alcohol cues such that individuals with the DRD4 VNTR > 7 had significantly greater mesocorticolimbic response to alcohol cues compared with individuals with < 7 repeats (Filbey et al. 2008b). In a recent study, a variation in the DRD4 was also found to be related to alcohol and inhibition phenotypes (Hack et al. 2010). It should be noted that this same study did not find any evidence for an association with DRD2. However, this study as well as other previous studies did not examine DRD2 rs1799732 (aka DRD2 −141C deletion), which has also been associated with alcohol dependence (Du & Wan 2009) (Prasad, Ambekar & Vaswani 2010). The functional significance of both the DRD2 rs1799732 and the DRD4 VNTR has been posited to be in diminished DA effects. In sum, these studies have implicated variants in DA genes in modulating craving and, more generally, in alcohol use disorders (AUD).

Given that neuro-biological models of addiction suggest that abnormalities exist in circuits that underlie craving (i.e. reward network; orbitofrontal cortex, striatum, ventral tegmental area) and response inhibition (i.e. reflective control network; prefrontal cortex, anterior cingulate gyrus) (Volkow, Fowler & Wang 2003), another pragmatic approach into elucidating the genetic contributions of addiction would be to investigate the mechanisms underlying impulsivity or disinhibition (Hack et al. 2010). Impulsivity is a multi-dimensional construct with several processes, including response inhibition or the ability to inhibit a pre-potent response (Dick et al. 2010; Lejuez et al. 2010). In a study by Bjork et al. (2004), deficits in impulsivity in alcohol abusing individuals was reported as increased rates of false alarms, discounting of delayed rewards for more immediate ones, increased rates of risky responses and higher levels of impulsivity and aggression. These deficits have been directly associated with alcohol consumption. For instance, in a group of alcohol-dependent individuals, poor behavioral inhibition and greater risk taking were related to greater quantity of consumption, and risk taking was associated with frequency of consumption (Weafer, Milich & Fillmore 2010). Impulsivity has also been reported to be associated with DRD2 and DRD4 polymorphisms. The DRD2 has been found to modulate performance in response inhibition tasks such as the Go/NoGo and the stop signal task (Enoch & Goldman 2001). Studies of the DRD4 have also found similar associations, although the direction of these associations has not been consistent. For example, in a group of healthy volunteers, it was reported that DRD4 variable number of tandem repeats (DRD4 VNTR) 7 repeat (DRD4.L) carriers have better performance on the Go/NoGo task, suggesting better ability to inhibit a motor response than non-DRD4 VNTR 7 allele repeat carriers (Kramer et al. 2009). Psychophysiological investigations associated this increased ability to inhibit with increased brain activation as measured by theta band response in prefrontal areas (Kramer et al. 2009). On the contrary, the DRD4 VNTR 7 allele has also been reported to be associated with lower response inhibition in stop signal task (Congdon, Lesch & Canli 2008) as well as greater risk taking and novelty seeking as measured by cognitive assessments and personality questionnaires (Eisenberg et al. 2007; Munafo et al. 2008; Roussos, Giakoumaki & Bitsios 2009). While the effects of DRD4 VNTR are not entirely consistent, the existing studies suggest that the DRD4 may play a role in response inhibition mechanisms. Lastly, a recent study found that both the DRD2 rs1799732 and DRD4 VNTR predicted 9–12% of the inter-individual variability in nucleus accumbens reactivity, which was associated with self-reported impulsive behavior. The authors suggested that the DRD2 and DRD4 polymorphisms contribute to variability in behavioral impulsivity and related risk for substance use disorders via modulations in the nucleus accumbens (Forbes et al. 2009).

In the current study, we examined the underlying neural mechanisms of the modulatory effects of DAergic genes on response inhibition. We focused on the DRD2 rs1799732 single nucleotide polymorphism (SNP) located in the intron region of DRD2 on CHR 11 and the DRD4 VNTR polymorphism in exon 3 of the gene coding for DA D₄ receptors. DRD2 rs1799732 has C or T alleles, which have also been reported as insertion (C)/(+) or deletion (−), respectively. The VNTR polymorphism varies between 2 and 11 repeats of a 48 bp coding region sequence, with a primarily trimodal distribution of 2, 4 and 7 repeat alleles (Ding et al. 2002). Because previous studies including our own have found that individuals with DRD2 rs1799732 C allele (a.k.a. DRD2 insertion variant) (Ishiguro et al. 1998; Prasad et al. 2010) and DRD4 VNTR > 7 repeats have greater sensitivity to the rewarding effects of alcohol (Franke et al. 2000; Filbey et al. 2008b), and because neuro-imaging studies suggest hypoactivation in regions underlying response inhibition in substance abusing populations (Kaufman et al. 2003), we predicted that individuals with the DRD2 rs1799732 C allele and DRD4 > 7 repeats will have less activation in these regions compared with their genetic counterparts.

MATERIALS AND METHODS

This study was carried out in accordance with the Declaration of Helsinki. The University of Colorado at Boulder IRB approved the use of human subjects for this study.

Participants

Fifty-three participants from the general population in the metro-Boulder/Denver area were recruited for this study and provided informed consent. A 12-hour abstinence from alcohol was required and verified by breathalyzer prior to the scan. The participants were a subset of larger sample that also participated in a study on alcohol cues previously reported (Filbey et al. 2008a). However, the Go/NoGo data are only available from this subset. Of note, recruitment and enrollment of this subset did not differ from the larger sample. All participants (mean age =22.5, 38 males) were right-handed heavy drinkers (mean drinks per drinking day, DDD = 5.2). Participants were genotyped for the DRD2 rs1799732 and DRD4 VNTR. Table 1 describes the demographic characteristics of the participants included in these analyses.

Table 1.

Demographic characteristics of the participants per genotype.

	DRD4 VNTR				DRD2 rs1799732
	DRD4.L		DRD4.S		C/C		T
	n = 23		n = 30		n = 39		n = 13
	Mean	SD	Mean	SD	Mean	SD	Mean	SD
Age	22.65	2.57	22.4	2.02	22.56	1.86	23.0	3.21
Range	21–33		21–28		21–28		21–33
	t = −0.451, P = 0.654				t = −1.027, P = 0.309
No. males	18	–	20	–	26	–	11	–
	χ2(1, n = 53) = 0.862, P = 0.353				χ2(1, n = 52) = 1.53, P = 0.216
Education	15.89	1.22	15.6	1.40	15.68	1.32	15.54	1.1
Range	14–19		12–20		12–20		14–17
	t = −0.880, P = 0.383				t = 0.347, P = 0.730
DDD	5.6	2.93	4.72	2.19	5.57	2.62	4.31	2.13
Range	2–13		2–10		2–13		2–10
	t = −1.1, P = 0.278				t = 1.56, P = 0.125
DDM	13.93	6	11.13	4.55	13.04	5.63	10	3.51
Range	6–25		3.5–20		3.5–25		4–16
	t = −1.9, P = 0.07				t = 1.84, P = 0.102
AUDIT	13.22	5.23	11.27	4.88	12.3	5	11.53	5.9
Range	5–24		3–22		5–23		3–24
	t = −1.399, P = 0.553				t = 0.465, P = 0.764

^*DRD2 rs1799732 genotype could not be determined for one male participant.

AUDIT = alcohol use disorders identification test; DDD = drinks per drinking day; DDM = drinks per drinking month.

DNA collection, extraction and storage

Participants were instructed to generate and deliver 5 ml of saliva in to a sterile 50-ml conical centrifuge tube. The saliva sample was then placed in the refrigerator, and lysis buffer was added within 24 hours. Tris-HCl, pH 8; EDTA, pH 8; SDS and NaCl were added at 100 mM, 20 mM, 0.5% and 125 mM final concentrations, respectively. The tubes were refrigerated until the DNA was extracted, usually within 48 hours. Proteinase K (0.2 mg/ml) was added, and the tubes were incubated at 65°C for 60 minutes. An equal volume of isopropyl alcohol was then added to each tube, the contents were mixed, and the DNA was collected by centrifugation at 3500 × g for 10 minutes. The DNA pellet was rinsed once with one ml of 50% isopropyl alcohol and allowed to air dry. For RNase treatment, 20 μg/ml RNAse A and 50 U/ml RNase T1 were added and incubated at 37°C for 30 minutes. To precipitate the DNA, two volumes of 95% ethanol was added and mixed by gentle inversion then collected by centrifugation at 3500 × g for 15 minutes. The samples were allowed to air dry followed by resuspension in 1 ml of 10 mM Tris-HCl, 10 mM EDTA buffer, pH 8.0, and placed in a 1.8-ml cryovial. The concentration of DNA is calculated from the absorbance at 260 nm analysis and then adjusted to a concentration of 10 ng/μL.

Genotyping

DRD2 rs1799732

Samples were genotyped using TaqMan^® primer and probe pairs; the probes were conjugated to two different dyes (one for each allelic variant). Custom probe/primer pairs were designed and ordered through Applied Bio-sytems (Life Technologies Corporation, Carlsbad, CA, USA). Primer and probe set for DRD2 rs1799732: Forward primer: 5′-CAAAACAAGGGATGGCGGAA; Reverse primer: 5′-CCACCAAAGGAGCTGTACCT; Probe 1: VIC-CTACCCGTTCCAGGCCG-TAMRA; Probe2: FAM-CTACCCGTTCAGGCCG-TAMRA.

The polymerase chain reaction (PCR) reaction mixture consisted of 20 ng of genomic DNA, 1× Universal PCR Master Mix, 900 nM of each primer and 200 nM of each probe in a 15-μl reaction volume. Amplification was performed using the TaqMan^® Universal Thermal Cycling Protocol and fluorescence intensity was measured using the ABI Prism 7500 Real-Time PCR System. Genotypes were acquired using the 7500 system’s allelic discrimination software (SDS version 1.2.3).

There were no DRD2 rs1799732 T/T individuals; therefore, the groups consisted of DRD2 rs1799732 C/C and DRD2 rs1799732 C/T individuals.

DRD4-VNTR polymorphism

The 48 base pairs (bp) repeat in exon 3 of the dopamine D4 receptor gene (DRD4) was genotyped using PCR, labeled primers and followed with fragment sizing on the Applied Biosystem’s 3100 Genetic Analyzer. This region was amplified using primers: DRD4-F 5′-VIC-GCTCA TGCTGCTGCTCTACTGGGC; DRD4-R 5′-CTGCGGGTCT GCGGTGGAGTCTGG; 2R = 279, 3R = 327, 4R = 375, 5R = 423, 6R = 471, 7R = 519, 8R = 567, 9R = 615.

Touchdown PCR was performed in a 20-μl volume containing 20 ng genomic template, 10 pmol of each primer (DRD4-F DRD4-R) and 200 μmol/l of each dNTP (ATP, CTP, TTP, 50% DEAZA GTD, 50% GTP). AmpliTaq Gold polymerase was used in ABI bufferII with 1.5 ml MgC12 and 10% dimethyl sulfoxide. After an initial denaturation of 10 minutes at 95°C, touchdown PCR was performed at 10 cycles of 95°C for 30 seconds, annealing starting at 65°C for 30 seconds and decrease 1°C every cycle, extension at 72°C for 1 minute, 30 seconds then 30 cycles of amplification at 95°C for 30 seconds, annealing at 55°C for 30 seconds, followed by annealing at 72°C for 1 minute, 30 seconds and a final extension step of 10 minutes at 72°C performed in an Eppendorf Mastercycler (Hauppauge, Suffolk, NY, USA) epgradientS thermal cycler. Fragment sizing was performed using an AB 3130 and analyzed with AB Peak Scanner software v1.0 (Life Technologies Corporation, Carlsbad, CA, USA).

Because only one individual was homozygous for the > 7 repeat allele, we combined DRD4 > 7 repeat heterozygotes and homozygotes into the DRD4 > 7 repeat allele group (DRD4.L) group.

Functional magnetic resonance imaging data acquisition

In order to investigate how DA genes modulate neural processes that underlie response inhibition, we collected blood oxygenated level dependent (BOLD) functional magnetic resonance imaging (fMRI) images of subjects while performing a cognitive task that investigates response inhibition. We used a previously described version of the Go/NoGo task (Garavan et al. 2002) in which participants were presented with four echo planar imaging (EPI) runs consisting of an alternating visual presentation of the letters X and Y (see Fig. 1).

Figure 1.

Illustration of the response inhibition paradigm.A version of the Go/NoGo task previously described by Garavan et al. (2002) was utilized to measure response inhibition. In this paradigm, each letter (‘X’ or ‘Y’) was presented for 700 ms followed by a fixed interstimulus interval (i.e. fixation cross) for 300 ms. The participants were instructed to use their right index finger to press a button following each alternating letter (e.g. ‘X’ ‘Y’ ‘X’; Go conditions) and to withhold their response if a break in the alternating pattern occurs (i.e. ‘X’ ‘Y’ ‘Y’; NoGo conditions). Each run had 312 stimuli presentations, 25 of which were NoGo stimuli that were presented in an unpredictable order to maintain response prepotency. A 14-second pre-run and an 8-second post-run saturation scan were also collected during each run resulting in a total scanning time of 334 seconds or 167 volumes per run

BOLD images were collected using a 3T GE Signa (Milwaukee, WI) scanner using a gradient echo, echo planar sequence (TR = 2000 ms, TE = 30 ms, Flip angle = 90, FOV = 24 cm, matrix size = 64 × 64, slice thickness = 5 mm3, number of slices = 29). Because response inhibition processes are frequently reported in the orbitofrontal cortex (OFC) and its surrounding areas, and because these areas are susceptible to severe signal dropout, we used a volume-selective z-shim EPI technique to acquire the functional images (Du et al. 2007). In this study, we acquired whole-brain fMRI scans with 29 slice locations parallel to the AC–PC line using a TR of 2 seconds. Z-shim compensation was applied in 5 out of the 29 slice locations, at the region including and immediately above the OFC. For a two-stage registration of the EPI images, a high-resolution T1-weighted 3D volume MRI (40 axial slices of part head, matrix = 512 × 512) acquired using the same slice angles, thickness and gap as the EPI images, and a high-resolution structural image using the inversion-recovery SPGR sequence (TI = 500 ms, FA = 7 degrees, slice thickness = 1.5 mm, 256 × 256 matrix, 200 mm × 200 mm FOV, bandwidth =15.6 kHz, 124 slices) were collected close to the EPI acquisition. A foam pillow was used to minimize head movement. A vitamin E capsule was placed on the right forehead as a fiduciary to mark the right hemisphere for orientation purposes. The tasks were presented using a goggle system, and responses were recorded using a fiber-optic response pad with four response buttons. Stimulus presentation was delivered using E-Prime (for visual presentations).

fMRI data analysis

Imaging data analyses were carried out using FEAT (FMRI Expert Analysis Tool) Version 5.43, part of FSL (FMRIB’s Software Library) using the following pre-statistics processing: non-brain removal using brain extraction tool (Smith 2002); spatial smoothing using a Gaussian kernel of full-width half-maximum (FWHM) 8 mm; mean-based intensity normalization of all volumes by the same factor; high-pass temporal filtering (Gaussian-weighted LSF straight line fitting, with sigma = 50.0 s). Time series statistical analysis was carried out using FILM with local autocorrelation correction (Woolrich et al. 2001). Regressors were created by convolving the stimulus timing files with a double gamma-variate hemodynamic response function. A multiple linear regression analysis was performed to estimate the hemodynamic response for different explanatory variables (i.e. successful Go trials, successful NoGo trials, unsuccessful NoGo trials) and a corresponding t-statistic indicates the significance of the activation of the stimulus. Contrast maps were created by contrasting (1) successful NoGo trials versus successful Go trials as a measure of ‘successful inhibition’; (2) unsuccessful NoGo trials versus successful NoGo trials as a measure of ‘inhibitory failure’; and (3) unsuccessful NoGo trials versus successful Go trials as another measure of failure to inhibit or simply, ‘false alarms’. Statistical maps were then registered to the Montreal Neurological Institute (MNI) template with a two-step process. First, EPI images were registered to the high-resolution MPRAGE image, which was subsequently registered to the 152 brain average MNI template. These registration steps were performed using FLIRT. After transformation of the masks into MNI space, higher-level analysis was carried out using FLAME (FMRIB’s Local Analysis of Mixed Effects). Group analyses were carried out using FLAME (Woolrich et al. 2004). To determine how the neural mechanisms that underlie response inhibition are related to drinking behavior, we performed correlation analyses between BOLD response and total alcohol use disorders identification test (AUDIT) score and DDD. We set our threshold and applied a type I error/ multiple comparison correction using FEAT’s cluster-thresholding method, which first defined contiguous clusters using a Z statistic maximum height threshold. Then, each cluster’s estimated significance level (from Gaussian random-field theory) was compared with the cluster probability threshold. Only clusters that met these two levels of threshold were considered significantly active.

RESULTS

The genotype groups did not differ in age, gender, education, drinks per drinking day or total AUDIT scores (see Table 1). T-tests of behavioral performance on the Go/NoGo task showed that the DRD4.S individuals performed significantly better than the DRD4.L individuals (P < 0.05). The DRD2 rs1799732 T group also performed better on the task than the C/C group (P < 0.01) (see Table 2).

Table 2.

Behavioral performance for each of the genotype groups on the Go/NoGo task.

DRD4 VNTR				DRD2 rs1799732
DRD4.L		DRD4.S		C/C		T
n = 23		n = 30		n = 40		n = 13
Mean	SD	Mean	SD	Mean	SD	Mean	SD
Successful Go Trials
1133.95	42.7	1122.64	62.4	1119.61	60.64	1149.31	22.51
t = 0.71, P = 0.48				t = −1.7, P = 0.1
Unsuccessful Go Trials
30.05	42.7	41.36	62.4	44.4	60.64	14.7	22.51
t = −0.7, P = 0.48				t = 1.71, P = 0.1
Successful NoGo Trials
37.1	12	44.7	13.8	38.8	14.3	48.8	8.52
t = 1.81, P = 0.038*				t = −30.05, P = 0.004**
Unsuccessful NoGo Trials
42.9	12.0	35.1	13.9	40.8	14.0	31.2	8.52
t = 2.21, P = 0.032*				t = 2.96, P = 0.006**

^*P < 0.05;

^**P < 0.01

Neural mechanisms of response inhibition and associations with drinking behavior

Looking at the entire sample (n = 53), we found that successful inhibition elicited significant widespread activation (volume = 88 163 voxels, max z = 7.9) that included areas previously reported in the literature during response inhibition such as the anterior cingulate gyrus (ACG), dorsolateral prefrontal cortex (DLPFC), medial prefrontal cortex (medPFC) and inferior frontal gyrus (IFG) in addition to mesocorticolimbic areas such as the OFC, striatum and ventral tegmental area (VTA) (cluster-corrected z = 1.9, P < 0.05). There were no significant correlations between BOLD response during successful inhibition and AUDIT or DDD.

Similar areas within the reflective control and reward networks (volume = 83,476 voxels, max z = 7.0) were also elicited during false alarms (cluster-corrected z = 1.9, P < 0.05). There were no significant correlations between BOLD response during false alarms and AUDIT or DDD.

DRD2 rs1799732

During successful inhibition, the T allele participants had significantly greater BOLD response in a cluster (volume = 7327 voxels, max z = 4.1) encompassing the right IFG, OFC and insula and middle and posterior cingulate gyrus compared with the C/C group (cluster-corrected z = 1.9, P < 0.05) (see Table 3 and Fig. 2). There was no difference between the two groups during inhibitory failure or false alarms, and the C/C individuals did not have greater activation than the T group in any of the comparisons. There were no significant correlations found between BOLD response and drinking behavior measures in any of the groups.

Table 3.

Loci of peak activation within significant clusters during successful inhibition (cluster-corrected P < 0.05, z = 1.9).

z-score	Localization	MNI x, y, z	BA
DRD2 rs1799732 (T > C/C)
Cluster size = 7327 voxels
4.10	L post-central gyrus	−20, −32, 56	3
3.69	R orbitofrontal cortex	48, 36, −22	47
3.53	R paracentral lobule	16, −30, 54	5
3.33	L pre-central gyrus	−28, −18, 54	4
3.33	R medial frontal gyrus	8, −4, 68	6
3.12	L pre-central gyrus	−22, −14, 68	6
DRD4 VNTR (DRD4.S > DRD4.L)
Cluster 1 size = 7310 voxels
3.81	L superior temporal gyrus	−50, −26, 4	22
3.48	L middle temporal gyrus	−34, −78, 30	19
3.26	L cerebellum	−28, −56, −28	-
3.21	L middle occipital gyrus	−38, −80, −8	18
3.08	L transverse temporal gyrus	−32, −32, 12	41
3.08	L middle temporal gyrus	−52, −40, −4	21
Cluster 2 size = 4451 voxels
3.34	R cingulate gyrus	16, −42, 34	31
3.03	L posterior cingulate	−16, −60, 18	31
2.87	R precuneus	20, −54, 16	31
2.81	R posterior cingulate	10, −62, 18	31
2.75	R postcentral gyrus	20, −38, 56	5
2.71	L posterior cingulate gyrus	−12, −40, 34	31
Cluster 3 size = 4445 voxels
3.75	L medial frontal gyrus	−14, 66, −12	11
3.66	L medial frontal gyrus	−12, 44, 18	9
3.47	L medial frontal gyrus	−18, 64, −12	10
3.45	L middle frontal gyrus	−22, 34, 26	9
3.45	L medial frontal gyrus	−8, 66, −8	10
3.43	L middle frontal gyrus	−40, 20, 32	9

Connectivity radius = 26 mm; R = right, L = left, BA = Brodmann area.

Figure 2.

Significant differences between genotype groups during response inhibition. During successful inhibition, between group differences were found for DRD2 rs1799732 such that T > C/C in the right IFG, and for the DRD4 VNTR such that the DRD4.S > DRD4.L in bilateral ACG, IFG and precuneus/posterior cingulate area (cluster-corrected P < 0.05, z = 1.9). The color scale represents z-scores, and the right side of the images represents the left brain hemisphere

DRD4 VNTR

During successful inhibition, the DRD4.S group participants had significantly greater BOLD response in three clusters compared with the DRD4.L group (cluster-corrected z = 1.9, P < 0.05). The first cluster (volume = 7310, max z = 3.81) included temporal areas such as the superior temporal gyrus. The second cluster (volume = 4451, max z = 3.34) included posterior areas, such as the precuneus and posterior cingulate gyrus. The third cluster (volume = 4445, max z = 3.76) included prefrontal areas such as the OFC, mid-frontal gyrus, ACG and IFG (see Table 3 and Fig. 2). There was no difference between the two groups during inhibitory failure or false alarms. DRD4.L individuals did not have greater activation than the DRD4.S group in any of the comparisons. There were no significant correlations found between BOLD response and drinking behavior measures.

Region of interest analyses

Region of interest (ROI) analyses were carried out by extracting the parameter estimates (PEs) for each of the ROIs found in Garavan et al. (2002) for the relevant contrasts successful inhibition, inhibitory failure or false alarms. We used the ‘Stops’ ROIs (i.e. middle frontal gyrus, precentral gyrus, cingulate gyrus, inferior parietal lobe, middle temporal gyrus, insula, thalamus, putamen) for the analysis of successful inhibition and the ‘Errors’ ROIs (i.e. middle frontal gyrus, inferior frontal gyrus, pre-central gyrus, cingulate gyrus, inferior parietal lobe, middle temporal gyrus, middle occipital gyrus, insula, putamen, caudate, thalamus) for the analysis of inhibitory failure and false alarms. Specifically, we used the center of activation for each ROI and then computed the necessary radius to generate the volume reported in the original paper. Spherical ROIs were then derived using the computed radius. For each ROI, average signal change across all voxels (no thresholding) was computed and entered into regression analyses. Analyses indicated that none of the ROIs showed significant difference between any of the SNPs reported here.

In addition, we used functionally-defined masks from our own analyses to examine whether the DRD2 and DRD4 polymorphisms were associated with differential neural response in task-relevant regions. Specifically, we used results from our analyses of the main effects of the task to mask the contrasts of the DRD2 and DRD4 SNPs. Thus, a mask was formed for each contrast of interest based on significant clusters in the group overall. Using the maps reduces the overall number of voxels being searched in the SNP analyses, thus reducing the total number of resels contained in the image, and reducing the overall correction for multiple comparisons. We found group differences during successful inhibition in the DRD2 rs1799732 SNP such that participants with a T allele showed greater activation compared to individuals with C/C alleles (cluster-corrected z = 1.9, P < 0.05) (volume = 13142, max z = 4.1) (see Fig. 3). BOLD response during inhibitory failure and false alarms showed no effects of either DRD2 or DRD4.

Figure 3.

Autosliced image of difference between DRD2 rs1799732 T > C/C during successful inhibition using a functionally defined mask of the NoGo correct > Go correct contrast (i.e. successful inhibition) (cluster-corrected P < 0.05, z = 1.9).The color scale represents z-scores, and the right side of the images represents the left brain hemisphere

DISCUSSION

In this study, we confirmed the hypothesis that variants in DA genes such as DRD2 rs1799732 and DRD4 VNTR are associated with differential patterns of neural response during response inhibition processes in heavy drinking adults. Other studies have used response inhibition tasks to demonstrate that loss of response inhibition predisposes individuals not only to initiation of use but also toward maintenance and relapse [e.g. Go/NoGo task (Saunders et al. 2008); Stroop task (Dao-Castellana et al. 1998); continuous performance test (Salgado et al. 2009); Hayling task (Noel et al. 2001)]. Taken together, evidence suggests that those who abuse alcohol have difficulty withholding a basic prepotent response. Our results confirmed our hypothesis that those with the risk alleles for alcohol dependence have less activation in areas reported to play a role in response inhibition. Specifically, we found that those with the DRD2 rs1799732 C/C and DRD4 VNTR > 7 repeats had less BOLD response during successful inhibition in expected areas such as the IFG, which parallels their behavioral response of decreased total number of successfully inhibited responses (compared to their genotype counterparts).

The association of DA genes with abnormal patterns of brain activity may indicate a dysregulation of the balance between impulsive and reflective systems (Bechara 2005). In Bechara’s model, prefrontal areas (medial prefrontal cortex and cingulate gyrus) are posited to underlie impulse and attention control, the insula is suggested to represent affective states. In substance-dependent populations, reduced activation in these regions has been posited to reflect compromised abilities to exert control over prepotent urges. Attenuation in these control areas (i.e. ACG and PFC) has been reported in cocaine users using the Go/NoGo task (Hester, Fassbender & Garavan 2004). Hypoactivation of an ‘error-related network’ that includes the ACG, pre-supplementary motor area (pre-SMA), insula, thalamus and parietal lobe, has also been reported in the addicted population during inhibitory failure (Garavan et al. 2002; Hester & Garavan 2004). For example, using a Go/NoGo task, it was found that compared to controls, opiate addicts had attenuated response in the ACG, which was associated with poorer behavioral task performance (Forman et al. 2004). Interestingly, in the control group, it was found that ACG activation during inhibitory failure positively predicted behavioral task performance suggesting that this error signal in the ACG or the ability to monitor errors may play a role in the loss of control in addiction.

Evidence for the functional significance of these SNPs has been reported in the literature. The DRD2 rs1799732 involves an insertion/deletion of a cytosine related to receptor density (Arinami et al. 1997). Specifically, the deletion (or T) allele is associated with a decrease in promoter strength as compared with the insertion (or C) allele. In schizophrenia, a decreased frequency of the deletion or T allele is thought to contribute to elevation of D2 receptor density. Similarly variations in DRD4 VNTR lengths have functional effects on D4 receptors (Asghari et al., 1995; Oak, Oldenhof & Van Tol 2000). For example, evidence shows that the > 7 repeat variant is associated with a blunted intracellular response to dopamine (Asghari et al., 1995) (Oak et al., 2000). In sum, the DRD2 rs1799732 C allele and DRD4 VNTR > 7 repeats are associated with attenuated dopamine receptor function. Moreover, D₂ and D₄ receptors are ubiquitous in brain networks (i.e. reward and response inhibition) that are important in addiction (Matsumoto et al. 1996). While D₂ receptors are more widely distributed in the striatum, D₄ receptors are more prevalent in the prefrontal cortex. Evidence suggests that the frontal cortex influences striatal activity through top–down corticostriatal projections by regulating DAergic input to the striatum (Roberts et al. 1994; Bertolino et al. 2000; Jackson et al. 2001). Thus, it is not surprising that prefrontal DAergic genes such as the DRD4 interact with striatal DAergic genes such as the DRD2. Neuro-computational models have begun to elucidate the important interplay between these regions and propose that frontostriatal projections allow the striatum to learn reinforcement probabilities and that PFC regulates this learning for control of behavior (Congdon & Canli 2005; Frank & Fossella 2010). Taken together, our findings of decreased BOLD response in inhibitory control areas during successful inhibition suggest the possibility that deficits in D₂ and D₄ receptor function may contribute to response inhibition in AUD.

While the link between impulsivity and addictive behavior is well established, only recently have the genetic determinants of impulsivity been a target of investigation. For instance, genetic variation in DA genes, such as the DAT and COMT, has been shown to have differential effects on neural response during inhibition processes (Congdon & Canli 2005; Congdon et al. 2008). DA genes have also been implicated in adaptive learning, which may account for error-related components of response inhibition. For example, in a neuro-computational investigation of the neural and genetic components of adaptive learning, Frank et al. (2007) reported on the associations of three striatal and prefrontal DA-regulating genes (Frank et al. 2007). Specifically, the authors reported that the DARPP-32 genotype predicted better reward learning, the DRD2 C957T SNP was associated with negative reinforcement learning, and, that the Val/Met polymorphism of the COMT gene modulated rapid adaptive behavior. These findings support dissociations of striatal and prefrontal DAergic learning mechanisms and separate gene effects on reinforcement learning parameters.

Interpretation of these findings is limited by the restricted age range of the sample (i.e. primarily college-aged adults) and the mild severity of alcohol abuse (i.e. five drinks per drinking day), which may decrease the generalizability of these findings. Future studies should include older adults in addition to individuals with greater severity of abuse and dependence. Of note, alcohol abuse or dependence and psychiatric comorbidities in addition to acute withdrawal effects were not formally assessed. Thus, characterization of how these mechanisms may be influenced by these disorders cannot be fully elucidated by the present study. However, because the present sample consisted mostly of heavy drinkers rather than alcohol-dependent individuals, withdrawal was unlikely to influence the results. Additionally, because of small sample sizes, these analyses did not investigate separate effects of the SNPs independent of the other. Thus, the challenge for future studies is the need to obtain sufficient sample sizes of allele-specific genotypic subgroups. Lastly, we report here that genetic factors influence neural mechanisms during successful inhibition; however, the specificity of the observed effects to lapses of inhibitory control could not be fully evaluated as we did not collect other measures of response inhibition performance (i.e. response times).

To conclude, the current findings indicate that fMRI BOLD response during response inhibition is an effective measure of the inhibitory control system and, could be a useful biomarker for genetic analyses of individual risk for AUD. In this report, we show evidence that DA genes modulate neural mechanisms that underlie response inhibition deficits associated with AUD. The importance of these genetic variants is that they result in differences in response inhibition (either behaviorally or neuronal), which increases the risk of alcohol abuse. The lack of difference in alcohol use despite these neural differences between genotype groups could be a reflection of the recruitment strategy of the study (i.e. heavy drinkers).