Studies of the 48 bp Repeat Polymorphism of theDRD4 Gene in Impulsive, Compulsive, AddictiveBehaviors: Tourette Syndrome, ADHD, PathologicalGambling, and Substance Abuse

David E. Comings,1* Nancy Gonzalez,1 Shijuan Wu,1 Radhika Gade,1 Donn Muhleman,1Gerard Saucier,2 Patrick Johnson,3 Ray Verde,3 Richard J. Rosenthal,4 Henry R. Lesieur,5Loreen J. Rugle,6 Warren B. Miller,7 and James P. MacMurray8

1Department of Medical Genetics, City of Hope Medical Center, Duarte, California2Department of Psychology, California State University, San Bernardino, California3Jerry L. Pettis V.A. Hospital, Loma Linda, California4Department of Psychiatry, University of California, Los Angeles, California5Institute for Problem Gambling, Pawtucket, Rhode Island6Trimeridian, Inc., Carmel, Indiana7Transnational Family Research Institute, Sunnyvale, California8Department of Psychiatry, Loma Linda University School of Medicine, Loma Linda, California

Prior studies have reported an associationbetween the presence of the 7 repeat alleleof the 48 bp repeat polymorphism of thethird cytoplasmic loop of the dopamine D4receptor gene (DRD4) and novelty seekingbehaviors, attention deficit hyperactivitydisorder (ADHD), Tourette syndrome (TS),pathological gambling, and substanceabuse. However, other studies have failed toreplicate some of these observations. To de-termine whether we could replicate theseassociations we genotyped 737 individualsfrom four different groups of control sub-jects, and 707 index subjects from four dif-ferent groups of impulsive, compulsive ad-dictive behaviors including substanceabuse, pathological gambling, TS, andADHD. Chi-square analysis of those carry-ing the 7 allele versus non-7 allele carrierswas not significant for any of the groups us-ing a Bonferroni corrected a of .0125. How-ever, chi-square analysis of those carryingany 5 to 8 allele versus noncarriers was sig-nificant for pathological gambling (p <.0001), ADHD (p < .01) and the total index

group (p < .0004). When the comparison in-cluded all 7 alleles the results were signifi-cant for gamblers (p < .0001), TS (p < .003),ADHD (p < .003), and the total group (p <.0002). There was a significant increase inthe frequency of heterozygosity versus ho-mozygosity for all alleles for pathologicalgamblers (p < .0031) and the total indexgroup (p < .0015), suggesting that heterosisplayed a role. In the substance abuse sub-jects a quantitative summary variable forthe severity of drug dependence, based onthe Addiction Severity Index, showed thatthe scores varied by increasing severityacross the following genotypes: 44 < hetero-zygotes < 77 < 22. Studies of other quanti-tative traits indicated an important role forthe 2 allele and the 22, 24, and 27 genotypes.All studies indicated that the role of theDRD4 gene in impulsive, compulsive, addic-tive behaviors is more complex than a solefocus on the 7 versus non-7 alleles. Am. J.Med. Genet. (Neuropsychiatr. Genet.) 88:358–368, 1999. © 1999 Wiley-Liss, Inc.

KEY WORDS: dopamine receptors; addic-tions; heterosis

INTRODUCTION

The most commonly studied polymorphism of the do-pamine D4 receptor gene (DRD4) is the 48 bp repeat inthe portion of the gene coding for the third intracyto-plasmic loop [Van Tol et al., 1992]. This results in three

Contract grant sponsor: National Institutes of Drug Abuse;Contract grant number: RO1-DA08417; Contract grant sponsor:Tobacco Related Research Disease Program; Contract grant num-ber: 4RT-0110; Contract grant sponsor: Gaming EntertainmentResearch and Education Foundation & National Center for Re-sponsible Gaming.

*Correspondence to: David E. Comings, Department of MedicalGenetics, City of Hope Medical Center, Duarte, CA 91010.

Received 16 January 1998; Revised 23 September 1998

American Journal of Medical Genetics (Neuropsychiatric Genetics) 88:358–368 (1999)

© 1999 Wiley-Liss, Inc.

major alleles representing 2, 4, and 7 repeats, and mi-nor alleles of 3, 5, 6, and 8 repeats. Expression studieshave shown that the major alleles differ moderately inthe interaction of the D4 receptor with the antipsychot-ic clozapine [Van Tol et al., 1992]. Despite the evidencefor a functional correlation with the major alleles, aswith many behavioral genetic investigations, the re-sults have been variable from study to study. Theseassociations cluster into a group of impulsive, compul-sive, and addictive behaviors, listed as follows.

Novelty Seeking

Ebstein et al. [1996] and Benjamin et al. [1996] in-dependently reported an association between the pres-ence of the 7 allele with higher novelty seeking scoresin individuals in the general population using Clon-inger’s Temperament and Character Inventory (TCI).Subsequent to this Malhotra et al. [1996] reported afailure to confirm this finding in Finnish subjects. Us-ing a different instrument, Jonsson et al. [1997] foundno association between the 7 allele and several person-ality traits in a Swedish population. Noble (personalcommunication) observed a modest association be-tween the 7 allele and novelty seeking behavior thatwas more significant when the association was madewith the presence of either the 7 allele of the DRD4gene and/or the Taq A1 allele of the DRD2 gene. Usingthe TCI, Ono et al. [1997] reported an association be-tween the 7 allele and novelty seeking in 153 normalJapanese students. Also using the TCI, Vandenberghet al. [1997] found no association between the 7 alleleand novelty seeking in older subjects, Pogue-Geile et al.[1998] found no association between with novelty seek-ing in 281 male and female twins, and Sander et al.[1997] found no association with either novelty seekingor alcoholism. Gelernter et al. [1997] found no associa-tion between the DRD4 gene and novelty seeking, test-ing several polymorphisms including the DRD4 7 al-lele.

Recognizing the discrepancy in results, Ebstein andBelmaker [1997] proposed a set of guidelines to maxi-mize the association with novelty seeking. These in-cluded always using Cloninger’s TPQ or TCI, or theNEO-PI-R, restricting studies the younger subjects (18to 32) with a high school or college education, indi-vidual analysis of men and women, and the exclusion ofindividuals with psychiatric disease. Using these crite-ria, Epstein et al. [1997] examined an additional 94subjects. Although there was no significant associationwith novelty seeking when grouped by the presence orabsence of the 7 allele, a significant difference in therange of novelty seeking scores was noted using a non-parametric test (P 4 .01). The effect was significant infemale but not male subjects. Epstein et al. [1997] con-cluded all the results combined support a modest rolefor the long alleles in novelty seeking at least in somepopulation groups.

Tourette Syndrome (TS) and Attention DeficitHyperactivity Disorder (ADHD)

Grice et al. [1996] reported an association betweenthe presence of the 7 allele and TS, using the transmis-

sion-disequilibrium test (TDT) in 64 family trios. How-ever, all but 12 of the trios came from four large fami-lies and the positive results came from only two ofthese families, suggesting genetic heterogeneity. InMexico, Nicolini et al. [1997] reported an associationbetween the 7 allele in 12 probands with chronic ticsand obsessive compulsive disorder (OCD) (TS + OCD)but not in 49 OCD probands without tics. While Barr etal. [1996] did not find evidence for linkage between theDRD4 gene and TS using either lod scores or nonpara-metric tests, these approaches may lack the power todetect the role of genes with small effects [Risch andMerikangas, 1996]. Hebebrand et al. [1997] evaluated102 TS probands and their parents using the TDT andfour different polymorphisms of the DRD4 gene thataltered the amino acid composition and had a fre-quency in Caucasians of >1%. All TDT tests were nega-tive including that for the 7 repeat allele in exon 3.Lahoste et al. [1996] reported an association betweenthe presence of the 7 allele and ADHD.

Pathological Gambling

Perez de Castro et al. [1997] reported an associationbetween the 7 alleles and pathological gambling . Theyexamined 68 Caucasian pathological gamblers and 68controls. The presence of the 7 allele was most commonin female pathological gamblers (p 4 .033).

Alcoholism

George et al. [1993] reported a significant increase inthe frequency of the 3 and 6 alleles in severe alcoholics.In a Japanese population, Muramatsu et al. [1966] ob-served a modest but significant increase in the fre-quency of the 5 allele in alcoholics that also carried thealdehyde dehydrogenase 2 gene. Adamson et al. [1995]found no association between any alleles or genotypesof the 48 bp repeat alleles in Finnish alcoholics versuscontrols, and Geijer et al. [1997] found no association inSwedish alcoholics. Parsian et al. [1997] found no sig-nificant differences in DRD4 allele frequencies be-tween 176 Caucasian controls versus 322 alcoholics,and no significant differences in transmitted alleles in58 family trios. Genotype frequencies were not re-ported. Others also found no association between theDRD4 gene and alcoholism [Chang et al., 1997].

Drug Dependence

Kotler et al. [1997] reported an association betweenthe 7 allele and opioid dependence. In Chinese subjects,Li et al. [1997] found a borderline association betweenthe 7 allele and heroin abuse (P 4 .07). When the longalleles (5–7) were compared with the short alleles (2–4), the association was just significant (P 4 .046, two-tailed). Mel et al. [see Ebstein et al., 1997] also foundan association between the DRD4 gene and heroinabuse in non-Ashkenazi Israelis.

Most studies of the possible association between the7 allele and schizophrenia [Barr et al., 1994; Hong etal., 1997; Kohn et al., 1997; Macciardi et al., 1994;Nanko et al., 1993; Petronis et al., 1995; Sommer et al.,1993; Tanaka et al., 1995] or manic depressive disorder

DRD4 in Impulse Disorders 359

[Lim et al., 1994; Perez de Castro et al., 1994], havebeen negative.

While such a history of successes and failures in theconfirmation of initial association studies is common inpsychiatric genetics, for the DRD4 gene it is reasonableto ask why. There are a number of possibilities. (1) Themost pessimistic is that the DRD4 gene is not associ-ated with any behavioral disorders and all the positiveresults are type I errors. However, it seems reasonableto conclude that with so many positive results some areprobably real. (2) Since these are polygenic disordersand since the effects of one gene such as the DRD4 geneare minor and account for only 1 to 2% of the variancefor the different phenotypes, for some populations theDRD4 gene may play a role in a given disorder or per-sonality variable while in different populations othergenes are more predominant and the association willbe negative. Thus, through a combination of populationheterogeneity and small effect size, variations fromstudy to study would be the expected outcome [Com-ings, 1998c]. (3) Examining subjects predominately bythe presence or absence of the 7 allele may not be themost sensitive approach and other strategies may bemore sensitive across multiple diagnoses and multiplesets of controls. (4) In a complex polygenic system, cer-tain alleles or genotypes may play a role in one pheno-type while other alleles or genotypes may play a role inother phenotypes. For example, in our studies of therole of many different genes in subjects with TS andADHD, specific genotypes of gene A and B may be ad-ditive for one phenotype but not additive, or even sub-tractive, in their effect on another phenotype. The 48bp polymorphism of the DRD4 gene is uniquely suitedto testing the possibility that different alleles or geno-types may have different effects in different pheno-types.

Over the past three years we have genotyped manysubjects for the 48 bp repeat of the DRD4 gene. Thepositive studies reviewed above suggest that the DRD4gene is involved in subjects that in general have im-pulsive, compulsive, and addictive behaviors. The posi-tive associations reported for the DRD2 gene have alsoinvolved impulsive, compulsive, and addictive behav-iors [Blum et al., 1995, 1996]. This is consistent withanimal studies showing an important role of dopaminein rewarding impulsive, hyperreactive, and stereo-typed behaviors [DiChiara et al., 1988; LeMoal and Si-mon, 1991]. Since most of the probands we have testedfall into this same category (TS, ADHD, pathologicalgambling, and substance abuse), and since we havealso examined multiple sets of controls, we felt thatthis wide range of individuals would allow us to exam-ine the following questions that may clarify the role ofthe DRD4 gene in behavior: (1) Is the increase in fre-quency of the 7 allele a characteristic of subjects acrossof wide range of impulsive, compulsive, and addictivebehaviors? (2) Is the presence or absence of the 7 allelethe best parameter to examine in studies of the DRD4gene or are other alleles or allele groupings, such as the5 to 8 alleles or the 2 allele, also associated with be-havioral disorders? (3) Does the phenomenon of mo-lecular heterosis, that we have observed for other do-pamine receptor genes (DRD1 [Comings et al., 1997a],

DRD2 [Comings and MacMurray, 1997a; Comings,1998b, 1998c], DRD3 [Comings et al., 1997a; Comingsand MacMurray, 1997a]), also occur at the DRD4 gene?(4) Are alleles or genotypes more important in identi-fying the role of the DRD4 gene in behavior? (5) Arethere differences in the effects of different alleles orgenotypes at the 48 bp polymorphism for different phe-notypes?

METHODSControl Subjects

In this study we used as controls four differentgroups of individuals with either no disorder or non-psychiatric disorders. These include the following. (1)One hundred and twenty-nine controls consisting ofstaff members, including both professional and nonpro-fessional personnel, from the City of Hope MedicalCenter and the Jerry L. Pettis Veterans Administra-tion Medical Center at Loma Linda, CA, screened toexclude drug and alcohol abuse/dependence [Comingset al., 1991]. These were termed COH controls. (2) Onehundred and thirty-six college students from the Cali-fornia State University at San Bernardino screened toexclude drug and alcohol abuse/dependence [Comingset al., 1997b; Comings et al., 1997a]. These weretermed SB controls. (3) Three hundred and sixty-eightrandom population-based volunteers for a study ofchildbearing motivation, performed in cooperation withDr. R. Miller. DNA from these subjects was obtainedfrom buccal smear kits mailed to the subjects alreadyinvolved in prior studies for the purpose of potentiallyidentifying genes associated with childbearing behav-iors [Miller et al., 1999b; Miller et al., 1999]. Thesewere termed the RM controls. (6) One hundred andfour subjects with multiple sclerosis from the Neuro-logical Disease Brain Bank (Dr. Wallace Tourtellotte,West Los Angeles V.A. Medical Center, Los Angeles,CA). These were termed the MS controls. In total therewere 737 nonpsychiatric non-Hispanic Caucasian con-trol subjects.

Index Subjects

We have used as index subjects four different groupsof individuals with a range of impulsive, addictive be-haviors, all from previous studies. (1) One hundred andsixty-five subjects from a prior study of the role of theDRD2 gene in pathological gambling [Comings et al.,1996b]. These were termed gamblers. (2) Two hundredand twenty-three subjects with TS from prior studies ofthe role of the DRD1, DRD2, DRD3, DBH, DAT1,TDO2, MAOA, and other genes in TS [Comings et al.,1991, 1993b, 1996a, 1996c, 1997a]. These were termedTS. (3) Fifty-two subjects with ADHD from prior stud-ies of the role of the DRD2 and TDO2 genes in ADHD[Comings et al., 1991, 1996a]. These were termedADHD. (4) Two hundred and sixty-seven subjects withalcohol and drug abuse/dependence from the AddictionTreatment Unit (ATU) of the Jerry L. Pettis VeteransAdministration Hospital that were part of earlier stud-ies of the molecular genetics of substance abuse [Com-ings et al., 1991, 1994, 1997a, 1997b; Gade et al., 1996;Johnson et al., 1997]. These subjects were obtained

360 Comings et al.

over a period from 1989 to 1997. The subjects obtainedsince 1994 were administered several psychologicaltests after they were thoroughly detoxified.

Quantitative Scores

The latter set of ATU subjects were administered theAddiction Severity Index, Fifth Edition [Hodgins andGuebaly, 1992]. This involved many different variablesto assess the severity of drug and alcohol abuse. Toreduce the problem of the analysis of multiple tests,these were summed to a quantitative ‘‘drug score’’[Comings et al., 1997b]. All TS subjects were given theHuman Behavioral Questionnaire. This questionnaire,based on the Diagnostic Interview Schedule [Robins etal., 1981] and Diagnostic and Statistical Manual ofMental Disorders (DSM-IIIR and DSM IV) criteria, al-lowed the compilation of 24 different quantitativetraits relating to a range of impulsive, compulsive, af-fective, cognitive, and other behaviors [Comings,1995b; Comings et al., 1996c]. To avoid the loss ofpower involved with multiple tests, we examined onlyfour of these variables: ADHD, oppositional defiant(ODD) and conduct disorder (CD), and tics [Comings,1995a]. The details of how these scores are compiledhave been provided elsewhere [Comings, 1995a; Com-ings et al., 1996c].

Personality Tests

All the SB subjects and all the ATU individuals ob-tained since 1994 were administered two personalitytests, the NEO Revised [Costa and McCrae, 1996], andthe TCI of Cloninger et al. [1993] and Svrakic et al.,[1993].

Allele Detection

The DRD4 48 bp repeat alleles were determined bythe polymerase chain reaction technique of Nanko etal., [1993]. We found that because of the GC rich natureof the polymorphic region, unless 5-azaguanadine com-pletely replaced guanadine, some heterozygotes weremissed, producing pseudohomozygosity. This requiredthe identification of the alleles by silver staining [Bu-dowle et al., 1991]. All the genotyping was done in theDepartment of Medical Genetics at the City of HopeMedical Center. All samples were identified only by arandom number.

Genotype Groupings

Several methods of genotype analysis were used. Foreach variable except D22-77, the genotype group antic-ipated to have the lowest scores was set to 0, and theother genotypes were scored as 1 or higher. These aresummarized below.

To emphasize the 7 allele, subjects were divided intothose carrying genotypes that contained the 7 allele(27, 47, 77) versus all other genotypes. For this vari-able, D7non7, those with non-7 genotypes were scored0 and those with a genotype containing the 7 allelewere scored 1. To emphasize the 44 genotype, subjectswere divided into those with the 44 genotype versus allothers. For this variable, D44non44, those with the 44genotype were scored 0, and those with all other geno-types were scored 1. To emphasize homozygosity ver-sus heterozygosity, subjects were divided into those ho-mozygous for any allele versus those heterozygous fordifferent alleles. For this variable, Dhomohet, thosewho were homozygotes (22, 33, 44, 55, 66, 77) werescored 0, and those who were heterozygotes werescored 1. One of us has suggested that the size of repeatalleles themselves may play a direct role in gene regu-lation [Comings, 1998a]. To test this hypothesis theD22–77 group scored genotypes by increasing length ofthe alleles. This involved the following scoring: 22 4 0,24 4 1, 44 4 2, 45–47 4 3, 77 4 4. Since these rep-resented the majority of subjects, and since it was notclear how to handle some of the minor genotypes suchas 26, 27, 36, and 37, these other genotypes were leftout of these analyses. While the use of multiple classi-fications of alleles and genotypes can be criticized onstatistical grounds [Goldman, 1996], a failure to exam-ine more than one classification also carries the risk ofmissing important phenotypic effects.

Statistics

The potential differences between the frequencies ofalleles or genotypes in the total control group versusthe individual or total subject group was examined bychi-square analysis using the asymptotic Pearson chi-square test of the statistical package StatXact3 forWindowst from Cytel Software Corporation (Cam-bridge, MA). A Bonferroni corrected a of .05/4 or .0125,was used. Analysis of variance was used to examine therelationship between DRD4 genotype and quantitativescores.

RESULTS

Table I summarizes the control and subject groups,the mean age, and the percentage of males in thesample. Clinical descriptions of the groups can befound in prior publications [Comings et al., 1991, 1944,1996b, 1996c; Miller et al., 1997a]. The most notabledifference between the control subjects versus the in-dex subjects was the sex ratio. Most of the index caseswere males, while the control cases tended to show aslight excess of females.

Total Allele Distributions

Table II lists the frequencies of the 2–8 repeat allelesin the four groups of control subjects, and four groups ofindex subjects with various impulsive and addictive be-haviors. The number and percentage of each allele inthe eight groups are shown. This allows an examina-tion of the variation in gene frequency that can occurwhen there is relatively small numbers subjects thecontrol groups. For example, the frequency of the 2

Variable 0 1+

D7non7 non 7 7D44non44 44 non-44Dhomohet homozygotes heterozygotesD22-77 22 1 4 24, 2 4 44,

3 4 45–47, 4 4 77D5–8 non 5–8 5–8

DRD4 in Impulse Disorders 361

allele ranged from 7.4 to 15.4% and the frequency of the4 allele ranged from 62.5 to 74.5% in the controls. All ofthe control groups were in Hardy-Weinberg equilib-rium. Since there were more males in the index groupsthan the control groups, it was reasonable to askwhether there was a significant difference in the dis-tribution of the alleles or genotypes in male versus fe-male controls. There were not. For alleles x2 4 4.75,d.f. 4 6, p 4 .57. For the genotypes x2 4 13.61, d.f. 416, p 4 .63.

All Alleles

When the total set of alleles was compared in a 2 × 7chi-square test of total controls versus index groups,the distributions were significant at a ø .0125 for thegamblers (p < .0001), TS (p ø .0003), ADHD (p ø.0003), and the total index group (p ø .0002).

7 Allele Frequencies

The frequencies of the 7 alleles in all groups areshown in Table IIIA. The frequencies of the 7 alleleranged from .087 to .190 in the control subjects andfrom .143 to .183 in the index subjects. The overallfrequency was .137 for the 737 total control subjectsand .166 for the 707 total index subjects. Individually

none of the subject groups were significant at a Bon-ferroni corrected a 4 .0125.

5–8 Allele Frequencies

Since many authors examine the effect of the DRD4gene by pooling the 5, 6, 7, and 8 alleles we also usedthis approach (Table IIIB). With a Bonferroni correcteda 4 .0125 the results were significant for the gamblers(P ø .0001), ADHD (ø .010), and the total group (P ø.0004).

Homozygosity Versus Heterozygosity

The frequencies of the genotype groups for the totalcontrol subjects and total index subjects are summa-rized in Figure 1. This showed a trend for index sub-jects to have a decrease in the frequency of homozygousgenotypes (22, 44) and an increase in heterozygousgenotypes involving the 5–7 alleles (45, 46, 47). Thissuggested that positive heterosis at the DRD4 geneswas occurring in subjects with impulsive and addictivebehaviors. To examine this the frequencies of homozy-gotes versus heterozygotes, and the homozygote/heterozygote ratios, are shown in Table IV. In the con-trol groups the homozygote/heterozygote ratios rangedfrom 1.091 to 1.419 with a mean of 1.213. In the subjectgroups the ratios ranged from .726 to 1.00 with a meanof .867. The ratios were significant using a Bonferronicorrected p of .0125 for the gamblers (p ø .0031) andthe total group (.0015).

Quantitative Traits

While the above results suggest that the 7 versusnon-7 allele comparison is not the most powerful ap-proach to examining the phenotypic effect of this DRD4polymorphism, it does not address the relative power ofthe different genotype variables. To continue to explorethe possible role of heterosis, in the following figures,the genotypes are arranged into three homozygousgenotypes (22, 44, 77) interspersed with two groupingsof heterozygotes (23, 24) and (45, 46, 47). These resultsfor the drug score for Caucasians only and for the totalATU group (Caucasians + Blacks + Hispanics) areshown in Figure 2. There was a U-shaped distributionwith individuals with the 44 genotype having the low-est scores, the 23-24 and 45- 46- 47 heterozygotes

TABLE II. Allele Counts for the 48 bp DRD4 Repeat Polymorphism—N (%)

Group T

Alleles

7 8 x2* P2 3 4 5 6

ControlsCOH controls 258 19 (7.4) 8 (3.1) 175 (67.8) 3 (1.2) 3 (1.2) 49 (19.0) 1 (.4)SB controls 272 26 (9.6) 4 (1.5) 194 (71.3) 6 (2.2) 2 (.7) 40 (14.7) 0 (0)RM controls 736 113 (15.4) 39 (5.3) 460 (62.5) 17 (2.3) 8 (1.1) 95 12.9) 4 (.5)MS controls 208 23 (11.0) 8 (3.8) 155 (74.5) 1 (.5) 1 (.5) 18 (8.7) 2 (1)Total 1474 181 (12.3) 59 (4.0) 984 (66.8) 27 (1.8) 14 (.9) 202 (13.7) 7 (.5)

SubjectsGamblers 330 20 (6.1) 7 (2.1) 214 (64.8) 14 (4.2) 8 (2.4) 60 (18.2) 7 (2.1) 36.96 <.0001Tourette synd 446 35 (7.8) 15 (3.4) 311 (69.7) 3 (.7) 16 (3.6) 64 (14.3) 2 (.4) 25.09 .0003ADHD 104 11 (10.6) 9 (8.7) 56 (53.8) 8 (7.7) 0 (.00) 19 (18.3) 1 (1) 25.50 .0003ATU 534 57 (10.7) 35 (6.6) 329 (61.6) 10 (1.9) 10 (1.9) 92 (17.2) 1 (.2) 14.83 .021Total 1414 123 (8.7) 66 (4.7) 910 (64.4) 35 (2.5) 34 (2.4) 235 (16.6) 11 (.8) 25.85 .0002

*Chi-square analysis versus the total controls, d.f. 4 6. Abbreviations as in Table I.

TABLE I. Summary of Groups Tested for DRD4 Alleles

Groups* N Mean age (S.D.) % males

Control subjects:COH 129 46.3 (15.3) 42SB 136 34.2 (7.9) 52RM 368 30.7 (4.3) 49MS 104 57.5 (13.0) 36Total 737

Index subjects:Gamblers 165 43.9 (10.1) 77TS 223 18.2 (13.4) 78ADHD 52 14.9 (12.2) 74ATU 267 41.5 (7.5) 99Total 707

*COH, City of Hope Medical Center and Jerry L. Pettis V.A. Medical Cen-ter Staff; SB, students from California State University at San Bernadino;RM, volunteers working with Dr. R. Miller; MS, subjects with multiplesclerosis; TS, Tourette’s syndrome; ADHD, attention deficit hyperactivitydisorder; ATU, Addiction Treatment Unit.

362 Comings et al.

higher scores, and the 22 and 77 homozygotes the high-est scores. Since the distribution for the total group (allraces) was virtually identical and had more power, itwas included and was significant at P 4 .0064

Personality Traits

A multivariate analysis of variance (MANOVA) ofthe five subscales of the NEO personality inventory

was performed using the D22-77 variable. This vari-able was chosen because it provided an assessment of awider range of genotypes than the variables that sim-ply divided the genotypes into two groups. The resultsare shown in Table V. The most significant associationwas with openness (P 4 .006). The total MANOVA wassignificant (p 4 .001). The distribution of the scores foropenness and conscientiousness across the versus ge-notype groups is shown in Figure 3. For both, thescores were lowest for those carrying the 22 genotype.Note that for these two scores, individuals with lowvalues are considered to have the most severe psycho-pathology. Since the scores for openness increased andstayed high for the remaining genotypes, the result forthe D22-77 scoring variable was significant. By con-trast the scores for conscientiousness peaked for the 44genotype then decreased across the remaining geno-types, and the result for the D22-77 scoring was of bor-derline significance. This again illustrates that to as-sess fully the association between DRD4 genotypes anddifferent phenotypes may require more complex assess-ment that specifically includes the 2 alleles and 22, 23,and 24 genotypes. By MANOVA, none of the seven fac-tors of the TCI, including novelty seeking, were signifi-cant using any of the DRD4 variables.

Quantitative Traits in TS and ADHD Subjects

To examine further the relationship between DRD4genotypes and quantitative traits, a totally differentset of subjects was examined, the COH controls, TS,and ADHD subjects. When all those that had com-pleted the Human Behavioral Questionnaire and hadDRD4 genotyping were included, the total n 4 362. To

Fig. 1. Frequency distribution of the major genotypes of the 48 bp re-peat polymorphism of the DRD4 gene in control subjects versus indexsubjects.

TABLE III. Frequency of the 7 Allele or 5 to 8 Alleles in Index Groups Compared With Controls

A. 7 versus non-7 allele for the 48 bp DRD4 repeat polymorphism7 Freq. Non-7 Total x2 P*

Control subjectsCOH 49 .190 209 258SB 40 .147 232 272RM 95 .129 641 736MS 18 .087 190 208Total 202 .137 1272 1474

Index subjectsGambling 60 .182 270 330 4.35 .037TS 64 .143 382 446 .11 .729ADHD 19 .183 85 104 1.68 .194ATU 92 .172 442 534 3.89 .048Total 235 .166 1179 1414 4.77 .028

B. 2–4 versus 5–8 alleles of the 48 bp DRD4 repeat polymorphismAny 5–8 Freq. Non 5–8 Total x2* P

Control subjectsCOH 56 .217 209 258SB 48 .176 224 272RM 124 .168 612 736MS 22 .106 186 208Total 250 .170 1224 1474

Index subjectsGambler 89 .270 241 330 17.70 <.0001TS 85 .191 361 446 1.04 .306ADHD 28 .269 76 104 6.64 .010ATU 113 .212 421 534 4.67 .035Total 315 .223 1099 1414 12.96 .0004

*One-way chi-square analysis versus the total controls, d.f. 4 1. Abbreviations as in Table I.

DRD4 in Impulse Disorders 363

minimize the problem of multiple comparisons, onlythose scores were chosen that were most relevant to theresults in the literature, i.e., the ADHD, ODD, CD, andtic scores. In previous studies [Comings et al., 1996c]we have found that using the entire range of quantita-tive trait variables can be a powerful method of exam-ining the effects of genes on behavior. Since both con-trols and TS and ADHD subjects completed the samestructured questionnaire, since the range of scores inthese groups tended to overlap, and since a broaderrange of scores provided more power, the scores forcontrols and TS and ADHD subjects were combined.These results are shown in Figure 4. The 22 genotypeoften had the highest scores whereas the 77 homozy-gotes often had the lowest scores. Using the D22-77variable, the only score that was significant using a

Bonferroni corrected a of .0125 was the conduct or CDscore (P 4 .006). This was significant because the scorefor the 22 genotype was one of the highest, and thescore for the 77 genotype was one of the lowest. For theADHD and ODD scores the higher scores were for the45–47 group, providing some support for positive het-erosis in that the scores were higher than for the 44 orthe 77 homozygotes. By contrast, the situation for the22, 23, and 24 genotypes was the opposite, providingsome support for negative heterosis in that the scorefor the 24 genotypes were lower than for the 22 or 44homozygotes.

DISCUSSION

Several conclusions are apparent from these studiesof four different control groups and four differentgroups of subjects with impulsive, addictive behaviors.

1. The division of the alleles into groupingsother than 7 non-7 may be superior. Benjamin etal. [1996] recommended a division of the alleles intoshort (2–4) versus long (5–8 repeats). For all four of ourindex groups and the total index group this gave moresignificant results. Thus, using an a of .0125 none ofthe groups were significant for the 7 versus non-7analysis while the gamblers, ADHD, and total groupswere significant using the 5–8 versus 2–4 analysis. Asdiscussed below, other methods of dividing the geno-types and alleles may also be relevant.

2. Heterosis also plays a role in the DRD4 gene.In regard to populations, heterosis refers to a situationwhere the progeny (hybrid) has a significantly greaterphenotypic effect than for either parental strain [Gard-ner and Snustad, 1981; Stern, 1973]. The vigor of hy-brid corn is a common example. We have used the term

TABLE V. Multivariate Analysis of Variance of the FiveFactors of the NEO Personality Inventory Versus the

D22-77 Variable (F and P)*

Agreeableness 2.86 .025Conscientiousness 2.47 .047Extraversion 0.97 .427Neurotocism 1.63 .168Openess 3.74 .006Total (Wilks) .001

*(N 4 167, Caucasians Only).

Fig. 2. Distribution by genotype group of the drug score. Solid line 4 allraces (Caucasians, Blacks, and Hispanics). Broken line 4 Caucasians only.( ) 4 number of cases.

TABLE IV. Homozygosity Versus Heterozygosity for the 48 bp DRD4 Repeat Polymorphism

Hom. Freq. Het. Freq.Hom/het

ratio x2* P

Control subjectsCOH 70 .543 59 .457 1.186SB 78 .574 58 .426 1.345RM 192 .522 176 .478 1.091MS 61 .587 43 .413 1.419Total 404 .548 333 .452 1.213

Index subjectsGamblers 69 .421 95 .579 0.726 8.73 .0031TS 111 .500 111 .500 1.00 1.59 .207ADHD 23 .423 28 .549 0.821 1.81 .178ATU 123 .464 142 .536 0.866 5.52 .019Total 326 .464 376 .536 0.867 10.10 .0015

*Chi-square analysis against the total of all controls, d.f. 4 1. Abbreviations as in Table I.

364 Comings et al.

molecular heterosis to refer to a situation where a poly-morphism for a single gene shows a significantlygreater effect in the heterozygote (usually 12) than ei-ther homozygote (11, 22). We have observed this phe-nomenon in many genes including other dopamine re-ceptor genes (DRD1, DRD2, and DRD3) [Comings,

1998b, 1998c; Comings et al., 1997a; Comings and Mac-Murray, 1997b; Lynch and Walsh, 1998]. This can beeither positive heterosis with heterozygotes havinghigher scores on quantitative traits than homozygotes,or negative heterosis with lower means in heterozy-gotes. This raises the possibility that heterosis mayplay a role in the phenotypic effect of all the dopaminereceptor genes. The observation of a significant in-crease in the frequency of the heterozygous versus ho-mozygous genotypes are summarized in Table IV. Thiswas significant with a 4 .0125 for the gamblers andthe total index groups. However, this conclusion is com-plicated by the finding that the quantitative scores forseverity of drug dependence (drug score) did not followthe pattern expected for heterosis, i.e., highest scoresfor the heterozygous genotypes (24, and 45 to 47), andlowest scores for the 22, 44, and 77 genotypes. Instead,the results for the drug score were consistent with arecessive effect, with the highest scores for the 22 and77 genotypes, or a codominant effect with higher scoresfor all non-44 genotypes.

3. Both positive and negative heterosis may oc-cur at the DRD4 gene. In our studies of the dopa-mine receptor genes heterosis was positive for theDRD2 gene [Comings, 1998b, 1998c; Comings and Mac-Murray, 1997b] and negative for the DRD1 [Comings etal., 1997a] and DRD3 [Comings et al., 1993a, 1996b]genes. One of the characteristics of heterosis was thatis could be phenotype specific and gender specific[Comings and MacMurray, submitted]. However, ineach of these cases and those in the literature, onlybi-allelic polymorphisms were studied. The more com-plex 48 bp polymorphism of the DRD4 gene, with 7alleles and multiple major genotype groups, allowedthe potential detection of more complex types of het-erosis. Thus, the ADHD and ODD scores (Fig. 4) wereconsistent with positive heterosis for the longer alleleswith the scores for the 45 to 47 heterozygous genotypesbeing higher than for the 44 and 77 homozygous geno-types, and with negative heterosis for the shorter al-leles with the scores for the 23 and 24 genotypes beinglower than for the 22 and 44 homozygous genotypes.This clearly is a tentative conclusion that will requireadditional studies in other populations.

4. The shorter alleles of the 48 bp play a role inthe phenotypic effect of the DRD4 gene. The pre-sent studies show that both heterozygosity and homo-zygosity for the 2 repeat allele also play significantroles in determining the phenotypic effect. For ex-ample, the drug score and the alcohol score for all raceswas higher for 22 homozygotes than for 77 homozy-gotes, and the scores for the 23 and 24 heterozygoteswere comparable to those for the 45 to 47 heterozy-gotes. The NEO scores for conscientiousness and open-ness were lowest for the 22 homozygotes. Despite thesmall number of 22 homozygotes, the low value for thisgenotype was the major reason there was a significantassociation between the DRD4 gene and openness. Fi-nally, scores for the 22 homozygotes were highest forall four of the scores examined, ADHD, ODD, CD, andtics.

Fig. 3. Distribution of the openness and conscientiousness NEO scoresfor the SB and ATU Caucasian subjects.

Fig. 4. Distribution of the quantitative traits among control, Tourettesyndrome (TS) subjects and attention deficit hyperactivity (ADHD) sub-jects. ADHD 4 attention deficit hyperactivity score. ODD 4 oppositionaldefiant score. CD 4 conduct score. Tics 4 tics severity score. ( ) 4 numberof cases.

DRD4 in Impulse Disorders 365

5. Some phenotypes may show a U-shaped asso-ciation with the DRD4 genotypes. While it mayseem counterintuitive that some phenotypes such asthe drug score would show a U-shaped pattern, wehave observed a similar result with some other repeats.For example, using a variable number of tandem re-peats (VNTR) at the MAOA gene, we observed thatboth the drug score and the ADHD score showed thelowest scores for the most frequent median sized al-leles, and increased scores for the rarer shorter andlonger alleles [Gade et al., 1998]. This is understand-able if the repeats themselves play a role in gene regu-lation. One of us has previously postulated that micro-and minisatellites play such a direct role through theiraffect on the formation of Z-DNA [Comings, 1998a].Since the amount of Z-DNA formed is very sensitive tovariations in both sequence and the length, differentsized alleles, and alleles with different sequences,could have a significant affect on the phenotype. The 48bp DRD4 allele is known to show a significant degree ofmicroheterogeneity [Lichter et al., 1993]. Thus, muta-tions that produce alleles that are either larger orsmaller than the most common alleles may have simi-lar positive or negative affects on the phenotype.

For the quantitative traits, the genotype patternscould be classified into U-shaped (for the drug scoreand conscientiousness score), a mixed heterosis pattern(for the ADHD and ODD scores), and a 22 predominantpattern (for the openness, CD, and tics scores). Giventhe number of variables involved, the small percentageof the variance explained, the number of cases, varia-tions in other background genes, the diagnosis, thephenotype, and the ethnic admixture, these differentpatterns could be predominately due to chance. Alter-natively, the patterns might be diagnosis or phenotypespecific. Additional studies using the same techniquesof analysis will be required to resolve this question.

In summary, when the examination of the role of theDRD4 gene is limited to the frequency of the 7 allele,and when the effect of the other repeat alleles andgenotypes is ignored, a significant portion of the phe-notypic effect of the DRD4 gene may be missed.

REFERENCES

Adamson MD, Kennedy J, Petronis A, Dean M, Virkkunn M, Linnoila Mand Goldman D. 1995. DRD4 dopamine receptor genotype and CSFmonoamine metabolites in Finnish alcoholics and controls. Am J MedGen (Neuropsych Genet) 60:199–205.

Barr CL, Kennedy JL, Lichter JB, Vantol HHM, Wetterberg L, Livak KJ,Kidd KK. 1994. Alleles at the dopamine D-4 receptor locus do not con-tribute to the genetic susceptibility to schizophrenia in a large Swedishkindred. Am J Med Genet 48:218–222.

Barr CL, Wigg KG, Zovko E, Sandor P, Tsui L–C. 1996. No evidence for amajor gene effect of the dopamine D4 receptor gene in the susceptibilityto Gilles de la Tourette syndrome in five Canadian families. Am J MedGen (Neuropsych Genet) 67:301–305.

Benjamin J, Li L, Patterson C, Greenberg BD, Murphy DL, Hamer DH.1996. Population and familial association between the D4 dopaminereceptor gene and measures of novelty seeking. Nature Genet 12:81–84.

Blum K, Wood RC, Sheridan PJ, Chen T, Comings DE. 1995. Dopamine D2receptor gene variants: association and linkage studies in impulsive,addictive, and compulsive disorders. Pharmacogenetics 5:121–141.

Blum K, Cull JG, Braverman ER, Comings DE. 1996. Reward deficiencysyndrome. Am Sci 84:132–145.

Budowle B, Chakiaborty R, Giusti AM, Eisenberg AJ, Allen PC. 1991.Analysis of the VNTR locus D1S80 by the PCR following by high reso-lution PAGE. Am J Hum Genet 48:137–144.

Chang FM, Ko HC, Lu RB, Pakstis AJ, Kidd KK. 1997. The dopamine D4receptor gene (DRD4) is not associated with alcoholism in three Tai-wanese populations: six polymorphisms tested separately and as hap-lotypes. Biol Psychiatry 41:394–405.

Cloninger CR, Svrakic DM, Przybeck TR. 1993. A psychobiological model oftemperament and character. Arch Gen Psychiatry 50:975–990.

Comings DE, Comings BG, Muhleman D, Dietz G, Shahbahrami B, Tast D,Knell E, Kocsis P, Baumgarten R, Kovacs BW, Levy DL, Smith M,Kane JM, Lieberman JA, Klein DN, MacMurray J, Tosk J, Sverd J,Gysin R, Flanagan S. 1991. The dopamine D2 receptor locus as a modi-fying gene in neuropsychiatric disorders. J Am Med Assn 266:1793–1800.

Comings DE, Muhleman D, Dietz G, Dino M, Legro R, Gade R. 1993a.Association between Tourette’s syndrome and homozygosity at the do-pamine-D3 receptor gene. Lancet 341:906.

Comings DE, Muhleman D, Dietz G, Dino M, Legro R and Gade R. 1993b.Tourette’s syndrome and homozygosity for the dopamine D3 receptorgene—reply. Lancet 341:1483–1484.

Comings DE, Muhleman D, Ahn C, Gysin R, Flanagan SD. 1994. Thedopamine D2 receptor gene: a genetic risk factor in substance abuse.Drug Alcohol Dependence 34:175–180.

Comings DE. 1995a. The role of genetic factors in conduct disorder basedon studies of Tourette syndrome and ADHD probands and their rela-tives. J Dev Behav Pediatr 16:142–157.

Comings DE. 1995b. Tourette syndrome: a hereditary neuropsychiatricspectrum disorder. Ann Clin Psychiatry 6:235–247.

Comings DE, Muhleman D, Gade R, Chiu C, Wu H, Dietz G, Winn–DeanE, Ferry L, Rosenthal RJ, Lesieur HR, Rugle L, Sverd J, Johnson P,MacMurray JP. 1996a. Exon and intron mutations in the human tryp-tophan 2,3-dioxygenase gene and their potential association with To-urette syndrome, substance abuse, and other psychiatric disorders.Pharmacogenetics 6:307–318.

Comings DE, Rosenthal RJ, Lesieur HR, Rugle L, Muhleman D, Chiu C,Dietz G, Gade R. 1996b. A study of the dopamine D2 receptor gene inpathological gambling. Pharmacogenetics 6:223–234.

Comings DE, Wu H, Chiu C, Ring RH, Dietz G and Muhleman D. 1996c.Polygenic inheritance of Tourette syndrome, stuttering, ADHD, con-duct and oppositional defiant disorder: the additive and subtractiveeffect of the three dopaminergic genes—DRD2, D H and DAT1. Am JMed Gen (Neuropsych Genet) 67:264–288.

Comings DE, Gade R, Wu S, Chiu C, Dietz G, Muhleman D, Saucier G,Ferry L, Burchete R, Johnson P, Verde R, MacMurray JP. 1997a. Stud-ies of the potential role of the dopamine D1 receptor gene in addictivebehaviors. Mol Psychiatry 2:44–56.

Comings DE, MacMurray JM. 1997a. Molelcular heterosis: Implicationsfor psychiatric genetics. Am J Med Gen (Neuropsych Genet) 74:656.

Comings DE, MacMurray JM. 1997b. Molecular heterosis: Implications forcomplex inheritance. Am J Hum Genet 61:A196.

Comings DE, Muhleman D, Gade R, Johnson P, Verde R, Saucier G, Mac-Murray J. 1997b. Cannabinoid receptor gene (CNR1): association withIV drug use. Mol Psychiatry 2:161–168.

Comings DE. 1998a. Polygenic inheritance and micro/minisatellites. MolPsychiatry 3:21–31.

Comings DE. 1998b. Molecular heterosis as the explanation for the con-troversy about the effect of the DRD2 gene on dopamine D2 receptordensity. Mol Psychiatry (in press)

Comings DE. 1998c. Why Different Rules are Required for Polygenic In-heritance: Lessons from Studies of the DRD2 Gene. Alcohol 16:61–70.

Costa PTJ, McCrae RR. 1996. The revised NEO personality inventory(NEO-PI-R). New York: Plenum.

de Castro IP, Torres P, Fernandez-Piqueras J, Saiz-Ruiz J, Llinares C.1994. No assciation between dopamine D4 receptor polymorphism andmanic depressive illness. J Med Genet 31:897–898.

366 Comings et al.

DiChiara G and Imperato A. 1988. Drugs abused by humans preferentiallyincrease synaptic dopamine concentrations in the mesolimbic system offreely moving rats. Proc Nat Acad Sci USA 85:5274–5278.

Ebstein RP, Novick O, Umansky R, Priel B, Osher Y, Blaine D, BennettER, Nemanov L, Katz M, Belmaker RH. 1996. Dopamine D4 receptor(D4DR) exon III polymorphism asociated with the human personalitytrait of novelty seeking. Nature Genet 12:78–80.

Ebstein RP, Belmaker RH. 1997. Saga of an adventure gene: Novelty seek-ing, substance abuse and the dopamine D4 receptor (DRD4) exon IIIrepeat polymorphism. Mol Psychiatry 2:381–384.

Epstein RP, Nemanov L, Klotz I, Gritsenko I, Belmaker RH. 1997. Addi-tional evidence for an association between the dopamine D4 receptor(DRD4) exon III repeat polymorphism and the human personalityatrait of novelty seeking. Mol Psychiatry 2:474–477.

Gade R, Muhleman D, MacMurray J, Comings D. 1996. Monoamine oxi-dase gene variants in Tourette syndrome and drug abuse. PsychiatGenet 6:168.

Gade R, Muhlemann D, MacMurray J, Comings DE. 1998. Correlation oflength of VNTR alleles at the X-linked MAOA gene and phenotypiceffect in Tourette syndrome and drug abuse. Mol Psychiatry 3:50–60.

Gardner, EJ, Snustad, DP. 1981. Principles of genetics, 6th ed. New York:John Wiley & Sons. p 544–547.

Geijer T, Jonsson E, Neiman J, Persson M-L, Brene S, Gyllander A, SedvallG, Rydberg U, Wasserman D, Terenius L. 1997. Tyrosine hydroxylaseand dopamine D4 receptor allelic distribution in Scandinavian chronicalcoholics. Alcoholism: Clin Exp Res 21:35–39.

Gelernter J, Kranzler H, Coccaro E, Siever L, New A, Mulgrew CL. 1997.D4 dopamine-receptor (DRD4) alleles and novelty seeking in sub-stance-dependent, personality-disorder, and control subjects. Am JMed Genet 61:1144–1152.

George SR, Cheng R, Nguyen T, Isreal Y, O’Dowd BF. 1993. Polymor-phisms of the D4 dopamine receptor alleles in chronic alcoholism. Bio-chem Biophys Res Comm 196:107–114.

Goldman D. 1996. High anxiety. Science 274:1483.

Grice DE, Leckman JF, Pauls DL, Kurlan R, Kidd KK, Pakstis AJ, ChangFM, Buxbaum JD, Cohen DJ, Gelernter J. 1996. Linkage disequilib-rium of an allele at the dopamine D4 receptor locus with Tourette’ssyndrome by TDT. Am J Hum Genet 59:644–652.

Hebebrand J, Nothen MM, Ziegler A, Klug B, Neidt H, Eggermann K,Lehmkuhl G, Poustka F, Schmidt MH, Propping P, Remschmidt H.1997. Nonreplication of linkage disequilibrium between the dopamineD4 receptor locus and Tourette syndrome. Am J Hum Genet 61:238–239.

Hodgins DC, Guebaly N. 1992. More data on the Addiction Severity Index.Reliability and validity with the mentally ill subtance abuser. J NervMent Dis 180:197–201.

Hong C-J, Lee Y-L, Sim C-B, Hwu H-G. 1997. Dopamine D4 receptor vari-ants in Chinese sporadic and familial schizophrenics. Am J Med Gen(Neuropsych Genet) 74:412–415.

Johnson JP, Muhleman D, MacMurray J, Gade R, Verde R, Ask M, KelleyJ, Comings DE. 1997. Association between the cannabinoid receptorgene (CNR1), and the P300 wave of event-related potentials, and drugdependence. Mol Psychiatry 2:169–171.

Jonsson EG, Nothen MM, Gustavsson JP, Neidt H, Brene S, Tylec A, Prop-ping P, Sedvall GC. 1997. Lack of evidence for alleleic association be-tween personality traits and the dopamine D4 receptor gene polymor-phisms. Am J Psychiatry 154:697–699.

Kohn Y, Ebstein RP, Heresco-Levy U, Shapira B, Nemanov L, Gritsenko I,Avnon M, Lerer B. 1997. Dopamine D4 receptor gene polymorphisms:Relation to ethnicity, no association with schizophrenia and responseto clozapine in Israel subjects. Eur Neuropsychopharmacol 7:39–43.

Kotler M, Cohen H, Segman R, Gritsenko I, Nemanov L, Lerer B, KramerI, Zer-Zion M, Kletz I, Ebstein RP. 1997. Excess dopamine D4 receptor(DRD4) exon III seven repeat allele in opioid dependent subjects. MolPsychiatry 2:251–254.

Lahoste GJ, Swanson JM, Wigal SB, Glabe C, Wigal T, King N, KennedyJL. 1996. Dopamine D4 receptor gene polymorphism is associated withattention deficit hyperactivity disorder. Mol Psychiatry 1:121–124.

LeMoal M, Simon H. 1991. Mesocorticolimbic dopaminergic network:functional and regulatory roles. Physiol Rev 71:155–234.

Li T, Xu K, Deng H, Cai G, Liu J, Liu X, Wang R, Xiang X, Zhao J, MurrayRM, Sham PC, Collier DA. 1997. Association analysis of the dopamineD4 gene exon III VNTR and heroin abuse in Chinese subjects. MolPsychiatry 2:413–416.

Lichter JB, Barr CL, Kennedy JL, Van Tol HHM, Kidd KK, Livak KJ.1993. A hypervariable segment in the human dopamine receptor(DRD4) gene. Hum Molec Genet 2:767–773.

Lim LCC, Nothen MM, Korner J, Rietschel M, Castle D, Hunt N, ProppingP, Murray R, Gill M. 1994. No evidence of association between dopa-mine D4 receptor variants and bipolar affective disorder. Am J MedGen (Neuropsych Genet) 54:259–263.

Lynch M, Walsh B. 1998. Genetics and analysis of quantitative traits.Sunderland, MA: Sinaure Associates, Inc. p 1–980.

Macciardi F, Verga M, Kennedy JL, Petronis A, Bersani G, Pancheri P,Smeraldi E. 1994. An association study between schizophrenia and thedopamine receptor genes DRD3 and DRD4 using haplotype relativerisk. Hum Hered 44:328–336.

Malhotra AK, Virkkunen M, Rooney W, Eggert M, Linnoila M, Goldman D.1996. The association between the dopamine D4 receptor (DRD4) 16amino acid repeat polymorphisms and novelty seeking. Mol Psychiatry1:388–391.

Miller W, Pasta D, MacMurray J, Comings DE. 1997a. Genetic influenceson childbearing motivation: further testing a theoretical framework.In: Genetic influences on fertility-related processes. NIH and Societyfor the Study of Social Biology, editors. Tuscon, AZ: Dec 3–5, 1997.

Miller WB, Pasta DJ, MacMurray J, Chiu C, Wu S, Comings DE. 1997b.Genetic influences on childbearing motivation: a theoretical frame-work and some empirical evidence. In: Advances in population: Psy-chosocial perspectives vol 3. Severy LJ and Miller WB, editors. London:Jessica Kingsley.

Miller WB, Pasta DJ, MacMurray J, Chiu C, Wu H, Comings DE. 1998.Dopamine receptor genes are associated with age at first intercourse.Soc Biol.

Muramatsu T, Higuchi S, Murayama M, Matsushita S, Hayashida M.1966. Association between alcoholism and the dopamine D4 receptorgene. J Med Genet 33:113–115.

Nanko S, Hattori M, Ikeda K, Sasaki T, Kazamatsuri H, Kuwata S. 1993.Dopamine D4 receptor polymorphism and schizophrenia. Lancet 341:689–690.

Nicolini H, Cruz C, Camarena B, P&aacute;ez F, Sidenberg D,De a FuenteJR. 1997. The 7-repeat variant of the DRD4 gene is increased in ob-sessive-compulsive behavior with tics. Biol Psychiatry 42:26S.

Ono Y, Manki H, Yoshimura K, Muramatsu T, Mizushima H, Higuchi S,Yagi G, Kanba S, Asai M. 1997. Association between dopamine D4receptor (D4DR) exon polymorphism and novelty seeking in Japanesesubjects. Am J Med Gen (Neuropsych Genet) 74:501–503.

Parsian A, Chakraverty S, Fisher L, Cloninger CR. 1997. No associationbetween polymorphisms in the human dopamine D3 and D4 receptorgenes and alcoholism. Am J Med Gen (Neuropsych Genet) 74:281–285.

Perez de Castro I, Ibanez A, Torres P, Saiz-Ruiz J, Fernandez-Piqueras J.1997. Genetic association study between pathological gambling and afunctional DNA polymorphism at the D4 receptor gene. Pharmacoge-netics 7:345–348.

Petronis A, Macciardi F, Athanassiades A, Paterson AD, Verga M, MeltzerHY, Cola P, Buchanan JA, Van Tol HHM, Kennedy JL. 1995. Associa-tion study between the dopamine D4 receptor gene and schizophrenia.Am J Med Gen (Neuropsych Genet) 60:452–455.

Pogue-Geile M, Ferrell R, Deka R, Debski T, Manuck S. 1998. Humannovelty-seeking personality traits and dopamine D4 receptor polymor-phisms: A twin and genetic association study. Am J Med Gen (Neuro-psych Genet) 81:44–48.

Risch N, Merikangas K. 1996. The future of genetic studies of complexhuman diseases. Science 273:1516–1517.

Robins LN, Helzer J, Croughan J, Ratclif KS. 1981. National Institute ofHealth diagnostic interview schedule. Arch Gen Psychiatry 38:381–389.

Sander T, Harms H, Dufeu P, Kuhn S, Rommelspacher H, Schmidt LG.1997. Dopamine D4 receptor exon III allels and variation of noveltyseeking in alcoholics. Am J Med Gen (Neuropsych Genet) 74:483–487.

DRD4 in Impulse Disorders 367

Sommer SS, Lind TJ, Heston LL, Sobell JL. 1993. Dopamine D4 receptorvariants in unrelated schizophrenic cases and controls. Am J Med Gen(Neuropsych Genet) 48:90–93.

Stern, C. 1973. Principles of human genetics. 3rd ed. San Francisco, CA:W.H. Freeman. p 455–458.

Svrakic DM, Whitehead C, Przybeck TR, Cloninger CR. 1993. Differentialdiagnosis of personality disorders by the seven-factor model of tem-perament and character. Arch Gen Psychiatry 50:991–999.

Tanaka T, Igarashi S, Onodera O, Tanaka H, Kameda K, Takahashi K,Tsuji S, Ihda S. 1995. Lack of association between dopamine D4 recep-

tor gene and schizophrenia. Am J Med Gen (Neuropsych Genet) 60:580–582.

Van Tol HHM, Wu CM, Guan H-C, Ohara K, Bunzow JR, Civelli O,Kennedy J, Seeman P, Niznik HB, Javanovic V. 1992. Multiple dopa-mine D4 receptor variants in human population. Nature 358:149–152.

Vandenbergh DJ, Zonderman AB, Wang J, Uhl GR, Costa PT Jr. 1997. Noassociation between novelty seeking and dopamine D4 receptor (D4DR)exon III seven repeat alleles in Baltimore longitudinal study of agingparticipants. Mol Psychiatry 2:417–419.

368 Comings et al.