ADHD and DAT1: Affected Family-Based Control Study using TDT Young Shin Kim, M.D., MPHDept. of EpidemiologyUC Berkeley

OutlineGenetics in Medicine: overviewAssociation StudiesADHD and DAT1: Affected Family-Based Control Study using TDTData Analysis

Genetics in Medicine; Two ApproachesGenetic EpidemiologyMacroscopic information (role of genetic and environmental factors, the familiarity of a disorder, and the mode of transmission) Molecular GeneticsFundamental information (identification of the normal products of the vulnerability genes, when and where the genes are normally expressed in the developing CNS, and insights into how specific alleles function to establish disease vulnerability) parametric methods - classical genetic linkage analysisnon-parametric methods - association studies, allele-sharing methods (affected sib-pair or affected relative studies)

Genetic Epidemiology (1)Twin study (initial indicator of importance of genetic vs. non-genetic factors)If MZ and DZ twins share their environment to the same extent, a higher concordance rate of disorders in MZ twins than in DZ twins suggests a strong genetic influence on the pathogenesis of the disorder Adoption Study (Useful for distinguishing genetic and environmental influences confounded by the shared environment in family studies) 1) a classic adoption study - compare the rates of disorders in biological relatives of the affected probands with those in the adoptive relatives2) an adoptees study to compare the rates of disorders in adopted offspring of affected parents with those in adopted offspring of unaffected parents3) a cross-fostering study to compare the rates of disorders in adopted offspring of affected parents who were raised by unaffected parents with those in adopted offspring of unaffected parents who were raised by affected parents

Genetic Epidemiology (2)Family Genetic Studies-Investigate the rates and patterns of occurrence of disorders in biological relatives of probands with a disorder -Useful tool for exploring the mode of transmission of a disorder -Caution; information bias, possibility of cultural transmission (phenotypes of interest are transmitted within families by non-genetic mechanism)

Segregation Analysis- Infer a best fitting mode of transmission of a disorder by ruling out those modes which do not fit the disorder- Visual inspection of obvious segregation patterns in pedigrees and performing formal statistical testing

Molecular Genetics (1)Studies of cytogenic abnormalities -Deletion and translocation of chromosomes -Clues to the chromosomal localization of disease vulnerability genes

Parametric studies -Classical genetic linkage analysis

Non-parametric studies -Association studies -Allele-sharing methods (affected sib-pair or affected relative studies)

Molecular Genetics (2)Linkage Analysis (powerful tool due to its ability to locate disease vulnerability genes precisely)

Attempt to identify disease vulnerability genes by investigating the association between the transmission pattern of a disorder in the pedigree and the linkage between a known genetic marker and putative genes thought to be responsible for disease, by detecting an RF smaller than 0.5 and estimating the magnitude of the linkage.A limitation of linkage analysis 1) disease model dependent (specifically, in psychiatric disorders, depends on correct assumptions regarding the involvement of a single major gene in the disease transmission, genetic homogeneity, and the precise mode of transmission) 2) statistical multiple testing3) limited applicability in disorders with complex traits, such as phenotypic variation, phenocopies, incomplete penetrance, genetic heterogeneity, polygenic inheritance and a high frequency of disease causing alleles in the population.

Molecular Genetics (4)Association studies

Detect a difference in allele frequencies at a particular locus, using a case-control study design. A positive allele association can indicate 1) the marker examined plays a causative role in the disorder 2) the marker is in linkage disequilibrium, and therefore, has no direct effect on the pathogenesis of the disorder but is closely linked to a disease-causing gene 3) it is a false positive finding due to either an artifact of population admixture (ethnic difference in allele frequencies) or polymorphism in the marker which leads to consequent multiple testing. Amendment for false positive findings 1) using a homogeneous population or using an internal comparison group, (affected family-based control) 2) Statistical adjustment for multiple testingThe advantages of association studies lack of a requirement for transmission models or assumptions and their potential to detect genes of small effect.

Molecular Genetics (5)A candidate gene approach

A special case of marker disease association study with a recombination fraction of 0 between marker and disease locusCandidate genes that are implicated in the pathogenesis of the disorder (e.g.; genes coding for specific NTs or receptors)

Molecular Genetics (6)The allele-sharing methods Affected sib-pair analysis, affected relatives analysis of a pedigreeDetect disease vulnerability genes: non-parametric method to detect linkage Investigators examine whether the inheritance pattern of a chromosomal region is inconsistent with random Mendelian segregation by showing that affected relatives or siblings inherit identical copies of the region more often than expected by chance

Association Studies (1)Linkage Equilibrium

Specific alleles at 2 loci; M, DM & D is in linkage equilibrium = P [MD] in gamete = P[M] * P[D] Deviation from this independent presence of M and D in haplotypes -> allelic association or linkage disequilibriumRecombination reduce the linkage disequilibrium by a factor of (1-)n (; recombination fraction probability of a recombination between 2 loci, n; generation) -> linkage disequilibrium can persist for hundreds of generations in tightly linked loci

Association Studies (2)Example; Ankylosing spondylitis (recessive gene)

At least 1 Allele at a marker loci B27(+)B27(-)Total

Disease D(+)72 3 75Status D(-) 372 75

Total7575150

RR (relative risk) = P[develop disease with risk factor]/P[develop disease without risk factor]= OR (in rare diseases) = 72*72/(3*3) = 576

Association Studies (3)Caution on interpretationAssociation can arise as an artifact due to population admixture (e.g.) association study between trait of the ability to eat with chopsticks and the HLA-A locus in the SF, then allele A1 would be positively associated because both the ability to use chopsticks and allele A1 is more frequent among Asians than Caucasians

Affected Family-based Control Association Studies (1)Basic Ideas (Rubinstein et al., 1981; Falk & Rubinstein, 1987)To avoid the difficulties in selecting appropriate control group, use parental data in place of non-related controls

Subjects Nuclear family with a single affected child -> type at the marker locus -> 2 parental alleles not transmitted to the affected child and they serve as a hypothetical control individual

Marker Genotypes in Nuclear Family (concept for AFBAC)M/mm/m

M/maffected child

hypothetical control; m/m

Affected Family-based Control Association Studies (2)Comparison with association studies

Disadvantages 1) more genotyping (2 in association studies, 3 in affected family-based control studies)2) Difficulty in sampling of trios

Advantagesovercoming problem of population stratification and false positive results of case-control studies

Affected Family-based Control Association Studies (3) Case-Control StudyGenotype M (+)M (-) Total

Case23 629Control 141529Total37 2158

AFBAC1 Haplotype Relative Risk (HRR) (1) Genotype-based Analysis (heterozygosity of allele M is not important) ; recessive model

Non-Transmitted genotype M (+)M (-) M/M, M/m m/mTotal

Transmitted M(+) 91423 (W)Genotype M(-) 5 1 6 (X)

Total14 (Y)15 (Z)29 (n)

each family contribute to one cell (total of 29 families)Assumption; distribution of marker genotypes from NT parental alleles in families is identical to the distribution of marker genotypes in the population.

AFBAC1 Haplotype Relative Risk (HRR) (2) HRR = W (M (+) affected child frequency) / X (M (-) affected child frequency)Y (M (+) in NT parental alleles) / Z (M (-) in NT parental alleles)= (W*Z)/ (X*Y) = (23*15)/(6*14) = 4.17 H0; No associationTest statistics; 2 test (14-5)2/(14+5), df=1, p=0.039 Similarity of HRR with RR in case-control studies = 0, HRR=RR (no recombination = linkage)

AFBAC2 Haplotype-Based Haplotype Relative Risk (HHRR) (1)Terlinger & Ott (1992)Compares observed frequencies of allele M in T (case) and NT (control) alleles (each parent has 2 allele, each family has 2 parents, thus total is 4n); unmatched analysis of TDT Parental alleleTotalMm

Transmitted52 658Non-transmitted391958Total9125116

AFBAC2 Haplotype-Based Haplotype Relative Risk (HHRR) (2)Assumption; 1) contribution of the parents are independent2) T and NT genotypes are independentH0; no association Test statistics; 2 test

AFBAC3 Transmission/Disequilibrium Test (TDT) (1)Spielman et al. (1993)

Non-Transmitted alleleTotalMm

TransmittedM331952m6 0 6Total391958

Classify each single parent according to his/her T and NT allele (Terwilliger & Ott; 1992)Each parent count per family-> 2n (29*2= 58); 58 parents in 29 families

AFBAC3 Transmission/Disequilibrium Test (TDT) (2)H0; no association and linkage (1-2) = 0 (=0; each heterozygous parent transmits its M allele to the affected child with probability of -> no association =1/2; no linkage)

Test statistic; McNemars 2 test = (b-c)2/(b+c) = (19-6)2/(19+6) = 6.76 (p=0.0093)Only heterozygous parents contribute to the analysis

Limitation TDT can detect linkage between the marker locus and the disease locus only if association (due to linkage disequilibrium) is present

DAT1 and ADHD: Affected Family-Based Control Study using TDTCharacteristics of Attention Deficit- Hyperactivity Disorder (ADHD)- Attention problem (inattention, distractibility)- Hyperactivity, impulsivityCommon; 3-6%Risk factor for antisocial and drug abuse in adulthood

Heritability of ADHDTwin studies 0.8 heritability estimatesFamily studiesmore ADHD in relatives of probands with ADHD compared to relatives of adoptive parents, normal controls, or psychiatric controlsSegregation analysis(1) autosomal dominant transmission with reduced penetrance of the hypothesized gene(2) single-gene effect vs. polygenic inheritance

DAT1 as a Candidate GeneDAT1 - VNTR (variable numbers of tandem repeats) of a 40 base-pair repeat sequence on chromosome 15.3- Majority; 10 repeats or 9 repeatsDAT1 and ADHD- Dopamine hypothesis; children with ADHD responds well to dopamine agonists and has inhibitory effects on the dopamine transporter (DAT1)- Animal studies1) overexpresesion of mutant rat dopamine transporter 2) mucous knockout studies

Aim of the StudyTo test the previously found family-based association of ADHD with the dopamine transporter 10-copy (480 bp) allele in this larger Korean sample

Study SubjectsChallenge; reduce phenocopy (someone plays the behavioral symptoms of ADHD without hypothesized genetics etiology)Inclusion criteria(a) Diagnosis of DSM-IV ADHD, Combined Type, determined by concurrence of two independent diagnosticians after review of all clinical information, corroborated by K-SADS and combined parent and teacher reports(b) Age between 6 and 12(c) WISC-III Full Scale IQ > 80(d) Participation of both biological parents(e) Parent informed consent and child assentExclusion criteria(a) Seizure disorder or neurological disease, bipolar mood disorder, pervasive developmental disorder, Tourette syndrome or chronic motor tic disorder, or uncorrected sensory impairment.(b) Parental history of bipolar mood disorder

Genetic Lab ProceduresDopamine transporter genotyping 1) PCR will be carried out in a 10 l volume containing 50 ng of genomic template, 0.5 M of each primer, one of which is 5' fluorescently labeled, 200 M of each dNTP (dATP, dCTP, dGTP, dTTP), 1 x PCR buffer, 2 mM MgCl2, and 0.5 units Taq polymerase (Amplitaq Gold). Samples will be amplified on a 9700 thermal cycler with an initial 12 minute step to heat-activate the enzyme, 40 cycles consisting of a denaturation step of 95 degrees C for 30 sec., an annealing step of 68 degrees C for 30 sec., and an extension step of 72 degrees C for 30 sec. 2) Products will be injected on an ABI 3700 multi-capillary array genetic analyzer with POP6 polymer. Products will be detected by laser-induced fluorescence using sheath flow on the ABI 3700. Electropherograms will be processed with Genescan software and alleles will be called with Genotyper software, blind to all but a number which is consecutively assigned and is not related to whether the subject is a child, father, or mother and without any indication of relationship to adjacent numbers.

Preliminary Analysis25 complete trios with ADHD combined type.Novel allele found; 365bp (7 repeat allele)Linkage Format

Family IDSubject IDDad IDMom IDGenderAffected StateAllele 01Allele 02483T483NTK02446200114444831K0242720021483483K02443946227212444483K02545800114835211K02527800214444831K02544445827812483483K02628500114444831K0264940021483483K02544928549412444483

2X2 Table Illustration (1)HRR AnalysisNon-Transmitted genotype 483 (+)483 (-) Total

Transmitted 483 (+)24125 (W)Genotype 483 (-)00 0 (X)

Total24 (Y)1 (Z)25 (n)

HRR = (25 * 1) / (0 * 1) = 2 test = not calculable due to 0s in cellsAccept null hypothesis; no association

2X2 Table Illustration (2)HHRR AnalysisParental alleleTotal483(+)483(-)

Transmitted43750Non-transmitted464450Total8911100

2 test = 0.919 (p=0.338)Accept null hypothesis; no association

2X2 Table Illustration (3)TDT Analysis

Non-Transmitted allele Total

483 (+)483 (-)

Transmitted483 (+)39443allele483 (-)7 0 7Total46450

McNemars 2 test with 1 d.f. = (4-7)2/(4+7) = 0.8182. (p>> 0.05)Accept null hypothesis; no association and no linkage

ImplicationsPossible that there is no association and linkage in Korean children with DAT1; difference with Caucasian children with ADHD

Limitations Small sample size and small number of heterozygous allele parents provided limited information on the TDT analysis

Analysis of more samples (aimed at least 120 trios) are underway

AcknowledgmentChild and Adolescent Psychiatry, University of ChicagoLaboratory of Developmental Neurosciences, Center for Developmental Disorders Bennett Leventhal, M.D.Ed Cook, M.D.Soo Jung Kim, M.D.Multi-center CollaborationKen Ah Cheoun, M.D. Yonsei University Medical CollegeBoo Nyun Kim, M.D., Seoul National University Medical CollegeHee Jung Yoo, M.D., Kyungsang National University Medical College

Thanks to the children and their families participated in this study

**An understanding of genetic influences on the pathogenesis of disorders is important for several reasons. Future interventions may target the disease vulnerability genes directly. Alternatively, as our model of disease pathogenesis becomes more refined and exact, it may be possible to develop novel treatments focused on the adverse downstream effects of these vulnerability genes. Advances in the knowledge of genetics can be divided into two broad categories - findings from genetic epidemiological approaches and molecular genetic studies. Genetic Epidemiology provides macroscopic information about a disorder, including the role of genetic and environmental factors on the pathogenesis of a disorder, the familiarity of a disorder, and the mode of transmission.The first step in molecular genetic studies is to identify the chromosomal regions where vulnerability genes are located. The analytic methods used to detect genetic linkage can be divided into parametric studies that use classical genetic linkage analysis, and non-parametric studies that include association studies and allele-sharing methods, such as affected sib-pair or affected relative studies. When specific genetic loci are identified as being associated with disease vulnerability, Molecular Genetics can provide fundamental information, including identification of the normal products of the vulnerability genes, when and where the genes are normally expressed in the developing CNS, and insights into how specific alleles function to establish disease vulnerability.We review several research methods that have proven useful in this field. *Twin study designs are the most useful single study design, providing the best initial indicator of the importance of genetic and non-genetic factors in the pathogenesis of a disorder. Classical twin studies compare the difference in concordance rates of a disorder between monozygotic (MZ) twin and dizygotic (DZ) twin pairs. This design is based on several assumptions including: 1) the classification of zygosity is accurate, and 2) MZ twins have the same genes whereas DZ twins, on average, share only half of their genes (Simonoff et al., 1994). If MZ and DZ twins share their environment to the same extent, a higher concordance rate of disorders in MZ twins than in DZ twins suggests a strong genetic influence on the pathogenesis of the disorder. Other variants of twin studies include studies of offsprings of twins, correlation studies between MZ twins, reared apart or together, and co-twin control studies. In particular, correlation studies and co-twin control studies examine differences in environmental exposure to explain the discordance of a disorder in MZ twin pairs. Adoption studies are particularly useful for distinguishing genetic and environmental influences that are confounded by the shared environment in family studies. Such designs include: 1) a classic adoption study to compare the rates of disorders in biological relatives of the affected probands with those in the adoptive relatives; 2) an adoptees study to compare the rates of disorders in adopted offspring of affected parents with those in adopted offspring of unaffected parents; and 3) a cross-fostering study to compare the rates of disorders in adopted offspring of affected parents who were raised by unaffected parents with those in adopted offspring of unaffected parents who were raised by affected parents (Stefanis et al., 1995). One practical and significant limitation of twin studies and adoption studies is the rarity and geographic dispersion of the study samples.

*Once twin and adoption studies have established that genetic factors play an important role in the pathogenesis of a disorder, family-genetic studies that investigate the rates and patterns of occurrence of disorders in biological relatives of probands with a disorder, are a very useful tool for exploring the mode of transmission of a disorder. They have also proven to be a valuable means for identifying related forms of psychopathology. However, family studies should be conducted carefully to prevent information bias. For example, information bias can be introduced easily when information is gathered retrospectively about all relevant family members based on information from a single source. In the absence of compelling twin or adoption data, attention should be paid to the possibility of cultural transmission, where phenotypes of interest are transmitted within families by non-genetic mechanism. Segregation analysis is a method used to infer a best fitting mode of transmission of a disorder by ruling out those modes which do not fit the disorder. It is conducted by, first, visual inspection of obvious segregation patterns in pedigrees and, second, performing formal statistical testing (Simonoff et al., 1994). The value of segregation analysis is for ruling out modes of transmission which do not fit a disorder. *Studies of cytogenic abnormalities, such as deletion and translocation of chromosomes, are useful for providing clues to the chromosomal localization of disease vulnerability genes.*Linkage analysis: Linkage of two genes at two adjacent loci indicates a failure to obey Mendels law of independent assortment, as the number of nonrecombinant offspring is greater than that of recombinant offspring (Simonoff et al., 1994). The closer the distance between the two genetic loci, the more likely that linkage is present. The recombination fraction (RF) is the ratio of recombinant offspring to the total number of offspring, *Association studies are a nonparametric method whose goal is to detect a difference in allele frequencies at a particular locus, using a case-control study design. A positive allele association can indicate: 1) the marker examined plays a causative role in the disorder under study; 2) the marker is in linkage disequilibrium, and therefore, has no direct effect on the pathogenesis of the disorder but is closely linked to a disease-causing gene; and 3) it is a false positive finding due to either an artifact of population admixture (ethnic difference in allele frequencies) or polymorphism in the marker which leads to consequent multiple testing. A false positive finding is the most serious limitation of association studies, but there are ways to minimize this problem. False positive findings due to population admixture can be prevented by using a homogeneous population or using an internal comparison group, such as an affected family-based control or haplotype relative risk method, and by confirming a tentative association by performing a transmission disequilibrium test (TDT) of a new sample from the same study population (Lander and Schork, 1994). Statistical adjustment can be used to reduce false positive findings caused by multiple testing, the other limitation of association studies. The advantages of association studies include their lack of a requirement for transmission models or assumptions and their potential to detect genes of small effect.*A candidate gene approach is useful for a focused search of disease vulnerability genes in molecular genetic studies. In this method, the genetic markers tested are not random but are found near candidate genes or are candidate genes that are implicated in the pathogenesis of the disorder, such as genes coding for specific neurotransmitters or receptors. However, given the large number of genes involved in CNS development, it is rare for this strategy to be successful. *The allele-sharing methods, which include affected sib-pair analysis or affected relatives analysis of a pedigree, are nonparametric method used to detect disease vulnerability genes (Lander and Schork, 1994). Contrary to linkage analysis, allele sharing methods are based on rejecting a model rather than constructing a model. Investigators examine whether the inheritance pattern of a chromosomal region is inconsistent with random Mendelian segregation by showing that affected relatives or siblings inherit identical copies of the region more often than expected by chance. The allele-sharing approach is more robust than linkage studies because it is not model-dependent, but it is less powerful than linkage analysis studies when correct assumptions are used.*Linkage EquilibriumSpecific alleles at 2 loci; M, DIf Probability of the haplotype MD (joint occurrence of alleles MD) in gamete = Probability (M) * Probability (D) -> M & D is in linkage equilibriumDeviation from this independent presence of Mand D in haplotypes -> allelic association or linkage disequilibriumIn evolutionary process mutation results in allele D at 2nd locus at the first locus for the first time. Some allele M is at the first locus in this gamete. -> then M & D are in complete linkage disequilibrium 2nd generationRecombination reduce the linkage disequilibrium; initial linkage disequilibrium is reduced by a factor of (1-)n (; recombination fraction probability of a recombination between 2 loci, n; generation) -> therefore, for tightly linked loci, linkage disequilibrium can persist for hundreds of generations*RR (relative risk) = P(develop disease with risk factor)/P(develop disease without risk factor)= OR = 72*72/(3*3) = 576*Caution on interpretationAssociation can arise as an artifact due to population admixture (e.g.; association study between trait of the ability to eat with chopsticks and the HLA-A locus in the SF, then allele A1 would be positively associated because both the ability to use chopsticks and allele A1 is more frequent among Asians than Caucasians)

*Basic Ideas;Rubinstein et al. (1981), Falk & Rubinstein (1987)To avoid the difficulties in selecting appropriate control group, use parental data in place of non-related controlsSubjects; Nuclear family with a dingle affected child -> type at the marker locus -> 2 alleles not transmitted to the affected child are combined more the marker genotype and serve as a hypothetical control individual*Comparison with association studiesDisadvantages more genotyping (2 in association studies, 3 in affected family-based control studiesDifficulty in sampling of triosAdvantagesovercoming problem of population stratification and false positive results of case-control studies*Genotype-based Analysis (heterozygosity of allele M is not important)

control caseeach family contribute to one cell (total of 29 families)Assumption; distribution of marker genotypes from NT parental alleles in families is identical to the distribution of marker genotypes in the population.*HRR = W (M (+) affected child frequency) / X (M (-) affected child frequency)Y (M (+) in NT parental alleles) / Z (M (-) in NT parental alleles)= (W*Z)/ (X*Y) = (23*15)/(6*14) = 4.17 H0; No associationTest statistics; McNemars 2 test (14-5)2/(14+5), df=1, p=0.039 Similarity of HRR with RR in case-control studies = 0, HRR=RR (no recombination = linkage) >0, NT parental genotypes cannot be regarded as a random sample from the general distribution. HRR is always at least as close to 1 as RR, thus even in the presence of recombination, HRR will not tend to overestimate the association between the marker and the disease.

Turning Point; Falk and Rubinstein proposed HRR to avoid incorrect conclusion from disease associations due to population stratification, but they did not focus on linkage. Otts analysis of the mathematical model for the HRR was the point of departure for the development of TDT.*Terlinger & Ott (1992)Compares observed frequencies of allele M in T (case) and NT (control) alleles (each parent has 2 allele, each family has 2 parents, thus total is 4n); unmatched analysis of TDT*Assumption; contribution of the parents are independentT and NT genotypes are independentH0; no association Test statistics; 2 test*Spielman et al. (1993)Classify each single parent according to his/her T and NT allele (Terwilliger & Ott; 1992)Each parent count per family-> 2n (29*2= 58); 58 parents in 29 families *H0; no association (=0; each heterozygous parent transmits its M allele to the affected child with probability of 1/2)H0; no linkage (=1/2)Test statistic No association; 2 test No linkage; McNemars 2 test = (19-6)2/(19+6) = 6.76 (p=0.0093)(Only heterozygous parents contribute to the analysis)*ADHD is characterized by symptoms of inattention and hyperactivity-impulsivity which are inconsistent with developmental level. DSM IV diagnostic criteria include Inattention; 1) fails to close attention to details, makes careless mistakes in schoolwork, work or other activities, 2) difficulty sustaining attentions, 3) often does not seem to listen when spoken to directly, 4) often does not allow through on instructions and fails to finish school work, chores, or duties in the workplace 5) often has difficulty organizing tasks and activities, 6) often avoids, dislikes, or is reluctant to engage in tasks that require sustained mental effort, 7) often lose things necessary for tasks or activities, 8) often easily distracted by extraneous stimuli and 9) often forgetful in daily activitiesHyperactivity; 1) often fidgets with hands or feet or squirm in seat, 2) often leaves seat in classroom or in other situations in which remaining seated is expected, 3) often runs around or climbs excessively in situations in which it is inappropriate, 4) often has difficulty playing or engaging in leisure activities quietly, 5) often on the go or acts as if driven by a motor, 6) talks excessivelyImpulsivity; 1) often blurts out answers before questions have been completed, 2) often has difficulties awaiting turn, 3) often interrupts or intrudes on others (butts into conversation or games)

It is a relatively common, but impairing, mental disorder, affecting 3-9% of school-aged children. ADHD often persists into adulthood and is a risk factor for development of antisocial and drug abuse disorders.

*Several types of study designs provide evidence for genetic factors in ADHD. First, twin studies are fairly consistent in reporting heritability of ADHD to be .50 or greater. Faraone and Biederman surveyed 11 twin studies and observed that heritability estimates averaged around .80 for ADHD and ADHD-related dimensions. Secondly, family studies find that ADHD runs in families. These studies report that ADHD is more prevalent among relatives of probands with ADHD, compared to relatives of adoptive parents, normals or psychiatric controls. Thirdly, segregation analysis has been used to compare models of familial transmission of ADHD. A segregation analysis of ADHD was consistent with autosomal dominant transmission with reduced penetrance of the hypothesized gene. In another study, this group examined the co-segregation of ADHD and learning disabilities and found that these two disorders were transmitted independently (Faraone et al., 1993). Of interest, they found evidence of assortative mating in that there was a higher than expected rate of reading disorders in the spouses of ADHD parents. Several other studies also favored a single-gene effect (Deutsch et al., 1990; Hess et al., 1995; Bailey et al., 1997), but polygenic inheritance has also been reported (Morrison and Stewart, 1974a). Thus, a variety of studies suggest that ADHD is highly heritable. Recent studies from our group indicate that even the comorbidities associated with ADHD aggregate in families in heritable patterns (Pfiffner et al., 1999). *Different forms of the DAT1 gene are characterized by variable numbers of tandem repeats (VNTR) of a 40 base-pair repeat sequence on chromosome 5p15.3. In the large majority of cases, these polymorphisms are either (a) 10 repeats (the most common), or (b) 9 repeats of the 40-bp sequence. Cook et al. (1995) reported a family-based association of the 10 copy (480-bp) DAT1 allele with child diagnostic status in a pooled sample of DSM-III-R ADHD and Undifferentiated Attention Deficit Disorder (UADD) in father-mother-child trios, based on Haplotype-based Haplotype Relative Risk (HHRR). The association remained statistically significant when non-Caucasians were excluded, as well as when UADD was excluded. It also remains significant if the TDT is applied instead of the HHRR (Cook et al., 1997). Similar results were reported by Gill et al. (1997), also using HHRR family-based analyses. Waldman et al. (Waldman et al., 1998) showed essentially this same phenomenon using the Transmission Disequilibrium Test (TDT), extending the findings by showing that linkage disequilibrium (a) was specific to the hyperactivity-impulsive (HI) (and not the inattention) features and (b) increased with severity of HI. Waldman et al. (Waldman et al., 1998) also showed that on average, siblings with more of the high-risk DAT1 alleles also had more ADHD symptoms, compared with siblings with fewer of the 10-copy alleles. In regression analyses across families, Waldman et al. (Waldman et al., 1998) showed that the number of DAT1 high-risk alleles was significantly related to the number of HI symptoms, even after controlling for major categories of ethnicity. Another replication of preferential transmission of the 10-copy in children with ADHD was reported by Daly et al. (Daly et al., 1999), detectable by either HHRR or TDT, and more clearly present in families with parents bearing a history of ADHD than in those with parents without such history.

Although the familial patterns and heritability of ADHD strongly suggest that genetic influences are at work, it is prudent to restrict the search to genes that are related in some way to the currently proposed pathophysiology of ADHD. The dopamine hypothesis has been advanced for many years, often in the context of a larger model involving both dopamine and norepinephrine (Pliszka et al., 1996). The most commonly used and well-studied treatment approach for ADHD is pharmacotherapy, which is effective, in the short run, for most children with ADHD. Dopamine agonists (including methylphenidate, dextroamphetamine, and pemoline) have been shown in numerous double-blind trials to be effective in the treatment of attentional dysfunction, hyperactivity, and impulsivity of ADHD (Casat et al., 1987; Zametkin and Rapoport, 1987; Greenhill, 1992; Hechtman, 1994; Rapport et al., 1994). Based on the inhibitory properties these agents have on the dopamine transporter (DAT1), the currently running study attempts to replicate and extend findings for DAT1s role as a primary candidate gene. This role is bolstered by findings that over expression of mutant rat dopamine transporter (with increased DA transport) leads to rapid habituation to a novel environment, increased sensitivity to locomotor effects of the psychostimulant, cocaine, and stronger place preference conditioning to cocaine (Miner et al., 1994). Further evidence comes from (a) mouse knockout studies showing that DAT knockouts are hyperactive (Giros et al., 1996) and hyperdopaminergic (Gainetdinov et al., 1999), with a loss of dopamine autoreceptor function (Jones et al., 1999); and (b) DAT functional research in adult ADHD (Dougherty et al., 1999), and association of the 9-copy allele with reduced DAT protein availability (Heinz et al., 2000). The activity of agents that are therapeutic for ADHD is not limited to DAT1, as these agents also promote membrane fusion of storage vesicles with subsequent release. Moreover, symptom improvement resulting from pre-synaptic activity does not rule out post-synaptic receptor characteristics from involvement in genetically transmitted susceptibility. Evidence that has accumulated since this study was launched encourages us to expand our investigation to other genes that may influence dopaminergic activity, particularly those for the DRD4 and DRD5 receptors. Though reports of involvement of DRD2 are intriguing (Comings et al., 1996), the within-family evidence for this gene has not been compelling (Rowe et al., 1999). *To reliably ascertain the presence or absence of ADHD in an expanded sample of school-age children referred for ADHD, carefully screened to decrease heterogeneity by excluding environmental or biological mechanisms that may mimic idiopathic ADHD.To test the previously found family-based association of ADHD with the dopamine transporter 10-copy (480 bp) allele in this larger sample.*The main challenge for sample collection is to reduce the proportion of individuals in the sample whose ADHD symptoms are the result of a phenocopy. The phenocopy would be characterized as someone who displays the behavioral manifestations of ADHD without the hypothesized genetic etiology. This problem has plagued genetic studies of other complex medical and psychiatric disorders besides ADHD. It will increase false negative findings by decreasing sample homogeneity.*Dopamine transporter genotyping. DNA will be extracted from whole blood by extraction with a PureGene kit from 10 ml of whole blood. The remainder of the blood will be frozen for future extraction, if necessary. Genotyping will be performed in the following manner. PCR will be carried out in a 10 l volume containing 50 ng of genomic template, 0.5 M of each primer, one of which is 5' fluorescently labeled, 200 M of each dNTP (dATP, dCTP, dGTP, dTTP), 1 x PCR buffer, 2 mM MgCl2, and 0.5 units Taq polymerase (Amplitaq Gold). Samples will be amplified on a 9700 thermal cycler with an initial 12 minute step to heat-activate the enzyme, 40 cycles consisting of a denaturation step of 95 degrees C for 30 sec., an annealing step of 68 degrees C for 30 sec., and an extension step of 72 degrees C for 30 sec. Similar methods will be used to amplify fluorescently labeled products of DRD4 and DRD5 with different fluorescent dyes so that there will be excellent sizing throughout the full range of these markers. For the DBH single nucleotide polymorphism, we will use single-base extension with fluorescent polarization (FP-TDI, http://psychiatry.uchicago.edu/ldn) (Chen et al., 1999) or single-based extension with detection by capillary electrophoresis (PE Applied Biosystems SnaPshot kit).For length variants (DAT1, DRD4 and DRD5), 1 l of each product will be added to 12 l of deionized formamide and 0.5 l of a custom-made ROX labeled internal size standard which will provide good coverage in the range of DRD5 and extend long enough to pick up the longer DRD4 fragments. Products will be injected on an ABI 3700 multi-capillary array genetic analyzer with POP6 polymer. Products will be detected by laser-induced fluorescence using sheath flow on the ABI 3700. Electropherograms will be processed with Genescan software and alleles will be called with Genotyper software, blind to all but a number which is consecutively assigned and is not related to whether the subject is a child, father, or mother and without any indication of relationship to adjacent numbers. Our accuracy has been demonstrated for all forms of genotyping, including genome-screen panels in which up to 16 microsatellite markers are pooled in one injection and incompatibilities were less than 1% in relatively large pedigrees (Veenstra-VanderWeele et al., 1998) and most recently by blind sequencing and FP-TDI genotyping of 394 samples for a single nucleotide polymorphism (3/394 discrepancies).