E.“
an 33.3

’21 .h’ 4

data?!
a.
.2-

.15..
5:... 43:...

a 35....
hump... ..
41...... x.«.(

.r... .
4...: 11
.116! .4“
.....q......¢ ,. ..

a mu!
1...“. 3..

5:.

«Eat; v
. it. ,
311.4!

.33

.4...
@1

.43 ,

.7; 232...»?
. vi? 1. \t
y .3. i

 

 

a o o 5 LIBRARY
Michigan State
University

 

 

 

This is to certify that the
dissertation entitled

A GENETIC STUDY OF ATTENTION DEFICIT/
HYPERACTIVITY DISORDER: CANDIDATE GENE
ASSOCIATION STUDIES USING HAPLOTYPES

presented by

LEEYOUNG PARK

has been accepted towards fulfillment
of the requirements for the

Doctoral degree in Genetics

Major Professor’s Signature

/5 five” 2005’

C/Datg(

MSU is an Affirmative Action/Equal Opportunity Institution

PLACE IN RETURN Box to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 unwind-n15

A GENETIC STUDY OF ATTENTION-
DEFICIT/HYPERACTIVITY DISORDER: CANDIDATE
GENE ASSOCIATION STUDIES USING HAPLOTYPES

By

Leeyoung Park

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Genetics program

2005

ABSTRACT
A GENETIC STUDY OF ATTENTION-DEFICIT/HYPERACTIVITY DISORDER:
CANDIDATE GENE ASSOCIATION STUDIES USING HAPLOTYPES
By

Leeyoung Park

Attention deficit hyperactivity disorder (ADHD) is one of the most heritable
complex disorders. Even with its high heritability, genome-wide scans do not Show
consistent results and candidate gene approaches have not been replicated in many cases.
Such inconsistent results indicate the lack of a major gene effect, which reinforces the
multigenic nature of ADHD, suggesting contributions from a large number of genes. In
order to detect genetic contributions for mapping complex diseases, linkage disequilibrium
(LD) has been the focus of recent research. Haplotype association studies use haplotypes
that consist of several polymorphisms usually in linkage disequlibrium near the gene region,
and consistently Show better detection than single marker studies.

Through this thesis research, several important considerations in haplotype
association studies were recognized. Two LD measurements, D’ and r2, differ depending on
the relationship between polymorphisms, so it is critical to consider which combination of
polymorphisms best captures the existence of risk alleles. Another consideration is that
there may be several or more polymorphisms in a haplotype block that affect a phenotype
in either a causative or a protective way. The third distinct point is that the detection power
varies depending on the choice of association testing and the contribution of a
polymorphism to the disorder.

Three candidate genes, the dopamine transporter gene (SLC6A3), the dopamine D4

receptor gene (DRD4), and the az-noradrenergic receptor gene (ADE/12A), were selected
depending on the catecholamine pathway, which is suspected to play a role in modulating
the major psychopathology of ADHD. Recognizing the importance of phenotypes in
association studies, gender difference and refined phenotypes were also studied. For gender
difference, the data suggest that genetic susceptibility to ADHD is regulated differently in
girls and boys. This posits important differences in the genetic susceptibility of the nervous
system between genders, suggesting that the same polymorphism performs differently due
to gender differences in dosage sensitivity in the catecholamine system.

This study reveals the association between all three candidate genes and ADHD
supporting the catecholamine pathway as a main etiology. Through this research, possible
major reasons for difficulties in mapping complex traits are identified. Moreover, by adding
more clarification to the gender difference and phenotype of ADHD, this study provides a

basic starting point for understanding the genetic etiology ofADHD.

To all (especially Jiehyzm).

ACKNOWLEDGEMENT

First of all, I would like to thank my advisor, Dr. Karen Friderici, for giving me
an opportunity to pursue this thesis research and guiding me with patience. I appreciate
the secondary guidance from Dr. Joel Nigg as an excellent collaborator and the useful
help from Dr. Irwin Waldman as a statistical consultant. I indicate that part of this
dissertation has partially or fully edited by Dr. Nigg and/or Dr. Waldman in pages 18-20,
24-25, 38-40, and 58-59. I also appreciate the guidance from my committee members, Dr.
Susan Ewart, Dr. Vilma Yuzbasiyan-Gurkan, Dr. Goncalo Abecasis, and Dr. Marc
Breedlove. The lab environment was comfortable to conduct experiments with familiar
faces, Mei Zhu (and her daughter), Ellen Wilch, Becky Bedilu, Soumya Korrapati, Sainan
Wei, Kathy Nummy, Kyle Gobrogge, and others. I would like to say thank you for trying
to help me in this work to Kathy Nummy and Kyle Gobrogge. It was always fun to
encounter the people, Donna and others, on the fifth floor in Biomedical Physical
Sciences building. In the Genetics program, I greatly appreciate the help from Dr. Barbra
Sears and Jeannine Lee as the current director and the secretary. I also would like to
thank Dr. Helmut Bertrand as the previous director and Dr. Sarah Elsea as my previous
committee member. It was lucky for me to be friends with Rebecca E. Slager and
Christopher Vlangos. I also would like to thank the instructors for the excellent lectures
on critical information. I would like to say thank you for their friendship to JungAn,
Jaeyeon, Yekyong, KangAe, and others. Also, I always appreciate the trust that I have got
through all the processe from Sunhee, Jungsoon and others in Korea. Finally with my

heart, I appreciate the endless support and belief in me from my family.

TABLE OF CONTENTS

ABSTRACT ........................................................................................................................ ii
ACKNOWLEDGEMENT .................................................................................................. v
TABLE OF CONTENTS ................................................................................................... vi
LIST OF TABLES ............................................................................................................. ix
LIST OF FIGURES ........................................................................................................... xi
KEY TO ABBREVIATIONS ............................................................................................ xii
CHAPTER I ....................................................................................................................... 1
Background ......................................................................................................................... 1
Introduction ................................................................................................................... 1
Diagnosis and etiology of ADHD ................................................................................. 2
Mode of inheritance and genome wide scans ............................................................... 5
Neurobiology of ADHD ................................................................................................ 7
Candidate gene studies of ADHD ................................................................................. 8
Linkage disequilibrium and haplotype studies. .......................................................... 10
Phenotypic considerations. ......................................................................................... 12
Study Design ............................................................................................................... 16
Chapter 2 ........................................................................................................................... l7
Candidate genes and initial analysis ................................................................................. 17
Introduction ................................................................................................................. 1 7
Candidate genes .......................................................................................................... 17
(I-ZA adrenergic receptor gene (ADRAZA) .......................................................... 17
Dopamine transporter gene (SLC6A3) ................................................................. 20
Dopamine receptor D4 gene (DRD4) ................................................................... 22

vi

Sample collection and demographic description ........................................................ 24

Initial association results ............................................................................................. 25
Initial haplotype analysis on ADRAZA ........................................................................ 3O
SNP selection ........................................................................................................ 3O
Linkage disequilibrium and haplotypes ................................................................ 33
Transmission disequilibrium testing using individual SNPS ................................ 35
Haplotype analysis ................................................................................................ 36
Discussion of initial haplotype analysis ...................................................................... 38
Conclusion .................................................................................................................. 40
Materials and methods ................................................................................................ 41
DNA Preparation ................................................................................................... 41
Genotyping ............................................................................................................ 43
Data Analysis ........................................................................................................ 46
Chapter 3 ........................................................................................................................... 48
Phenotypic Consideration ................................................................................................. 48
Introduction ................................................................................................................. 48
DSM IV based associations. ....................................................................................... 50
Endophenotypes .......................................................................................................... 54
Gender difference ........................................................................................................ 57
DRD4. ................................................................................................................... 59
SLC6A3. ............................................................................................................... 6O
ADRAZA. ............................................................................................................. 61
Discussion ............................................................................................................. 65
Conclusion .................................................................................................................. 69

vii

Chapter 4 ........................................................................................................................... 70

Haplotype Analysis ........................................................................................................... 70
Introduction ................................................................................................................. 70
Updated haplotype analysis for ADRAZA ................................................................... 71
Haplotype analysis of SLC6A3 ................................................................................... 77
Haplotype analysis of DRD4 ...................................................................................... 86
Conclusion .................................................................................................................. 92
Materials and methods ................................................................................................ 93

Setting up 384 well plates and quantification ....................................................... 93
TaqMan assays ...................................................................................................... 96
Sequencing ............................................................................................................ 96
Haplotype analysis ................................................................................................ 97

Chapter 5 ........................................................................................................................... 98

Discussion ......................................................................................................................... 98
Introduction ................................................................................................................. 98
LD and functional polymorphisms in haplotype analysis .......................................... 98
Differential detection of associations between TDT and case-control test ............... 101
Opposite direction of VNTR on SLC6A3 in our sample population ........................ 103
Possibility of genotyping error .................................................................................. 107
Future studies ............................................................................................................ 108

Literature Cited ................................................................................................................ 110

viii

LIST OF TABLES

Table 1. Diagnostic criteria for Attention-Deficit/Hyperactivity Disorderl4 ...................... 3
Table 2. Genotype Association Results (p-values of chi-square test). .............................. 26
Table 3. TDT results for each ADRAZA SNP and ADHD subtype. ................................. 28
Table 4. QTDT results for ADHD symptom dimensions .................................................. 29
Table 5. Polymorphism information for the ADRA2A locus ............................................. 32

Table 6. Transmission disequilibrium test results for each haplotype and ADHD subtype.
........................................................................................................................................... 36
Table 7. Haplotype analysis for ADRA2A using QTDT and ADHD symptom dimensions.
........................................................................................................................................... 37
Table 8. Genotype Association Results (p-values of chi-square test). .............................. 51
Table 9. Genotype Association Results (p-values of chi-square test on only Caucasians).
........................................................................................................................................... 52
Table 10. Transmission disequilibrium test results for each ADRAZA SNP and ADHD
subtype. ............................................................................................................................. 53
Table 11. QTDT results for ADHD symptom dimensions associated with ADRAZA
polymorphisms .................................................................................................................. 54

Table 12. Endophenotype analyses (p-values of tests for total evidence of association and

QTDT) ............................................................................................................................... 56
Table 13. Case-control association of DRD4 (controls vs ADHD all types). ................... 62
Table 14. Case-control association of SLC6A3 (controls vs ADHD all types). ................ 63
Table 15. Case-control association of ADRA 2A (controls vs ADHD all types) ................ 64
Table 16. SNP summary in ADRAZA. ............................................................................... 72
Table 17. Linkage disequilibrium in ADRAZA. ................................................................ 72

ix

Table 18. Case-control association tests for individual SNPS on ADRA2A. ..................... 74

Table 19. QTDT for individual SNPS on ADRA2A ........................................................... 74
Table 20. Haplotype association tests through PHASE. ................................................... 75
Table 21. Haplotyope association tests through UNPHASED. ........................................ 76
Table 22. Polymorphism summary in SLC6A3. ................................................................ 78
Table 23. Linkage disequilibrium in SLC6A3. .................................................................. 79

Table 24. Case-control association tests for individual polymorphisms on SLC6A3. ...... 81
Table 25. p-values of the case-control haplotype analyses using PHASE. ....................... 82
Table 26. Summary of haplotype frequencies with all polymorphisms ............................ 83

Table 27. Summary of haplotype frequencies with polymorphisms in the second LD

block .................................................................................................................................. 84
Table 28. p-values of the association tests from UNPHASED. ........................................ 85
Table 29. Haplotyope association tests through UNPHASED. ........................................ 85
Table 30. Summary of typed polymorphism in DRD4. .................................................... 87
Table 31. Linkage disequilibrium in DRD4. ..................................................................... 88
Table 32. The case-control associations of individual polymorphisms. ........................... 89
Table 33. Haplotype association studies using PHASE. ................................................... 91
Table 34. Possible haplotypes of three-marker model depending on LD. ...................... 100
Table 35. SLC6A3 expression assay depending on their VNTR genotypes .................... 103
Table 36. SC6A3 density depending on their VNTR genotypes. ................................... 104

LIST OF FIGURES

Figure 1. Catecholamine pathways and candidate genes. ................................................... 9
Figure 2. Schematic representation of dopamine transporter ........................................... 22
Figure 3. SNP location, linkage disequilibrium and haplotype distribution. .................... 34
Figure 4. Possible schematic diagram of synapses of girls and boys. .............................. 68
Figure 5. Graphical summary of LD in ADRAZA. ............................................................ 73
Figure 6. Gene structure and genotyped polymorphisms of SLC 6A 3. .............................. 78
Figure 7. Graphical summary of LD in SLC6A3 (D’). ..................................................... 80
Figure 8. Gene structure and typed polymorphisms. ........................................................ 86
Figure 9. Graphical summary of linkage disequilibrium in DRD4 (D’). .......................... 88
Figure 10. Standard curve using serial dilution of ABI control DNA. ............................. 94
Figure 11. Allelic discrimination plot of a test assay. ....................................................... 95
Figure 12. One possible description of haplotype structure using three SNPS. ................ 99
Figure 13. VNTR subtypes. ............................................................................................ 106
Figure 14. SLC6A 3 expression depending on VNTR subtypes. ..................................... 106

xi

ABI
ADHD
ADHD-C
ADHD-PI
ADRA2A
ANOVA
ASP

bp
['23I]B-CIT
DA

NE

DA T1
DBH
DISC-IV
1)sz
DRD4
DRD5
DSM-IV
DZ twin
EEG
ETDT

fl\/IRI

KEY TO ABBREVIATIONS

Applied Biosystems Inc.

attention deficit/ hyperactivity disorder
ADHD-combined type

ADHD-predominantly inattentive type
ot—2A-noradrenergic receptor gene

analysis of variance

affected sib pair

base pair

[1231] methyl 3B-(4-iodophenyl)tropane-2B-carboxylate
dopamine

noradrenergic

dopamine transporter gene

dopamine B-hydroxylase gene

the NIMH diagnostic interview Schedule for children, 4th edition
dopamine D2 receptor gene

dopamine D4 receptor gene

dopamine D5 receptor gene

diagnostic and statistical manual of mental disorders, 4”1 edition
dizygotic twin

electroencephalogram

extended transmission disequilibrium test

functional magnetic resonance imaging

xii

GABA
HRR

LD

LOD
MAP kinase
MLS
MZ twin
PC R
PKC
QTDT
QTL

RF LP
SC6A3
SLC6A3
SNP
SPECT
TDT
TSC
UTR

VNTR

gamma aminobutyric acid

haplotype relative risk

linkage disequilibrium

logarithm of odds ratio

mitogen activated protein kinase

maximum LOD score

monozygotic twin

polymerase chain reaction

protein kinase C

quantitative transmission disequilibrium test
quantitative trait loci

restriction fragment length polymorphism
sodium-dependent dopamine transporter
dopamine transporter gene

single nucleotide polymorphism

single photon emission computed tomography
transmission disequilibrium test

the SNP consortium

untranslated region

variable number of tandem repeats

xiii

CHAPTER 1

Background

Introduction

Attention Deficit Hyperactivity Disorder (ADHD) is a behavior disorder with
strong heritability (0.7) characterized by marked and pervasive inattention, hyperactivity
and impulsiveness resulting in impaired social and/or academic functioning'. 1t
commonly affects 5- 10% of children and adolescents and more than 3% of adults”. Boys
are affected 3-8 times more frequently than girls7. ADHD usually occurs in conjunction
with other major psychiatric disorders. Common comorbidity disorders and their relative
frequencies are as follows: oppositional defiant disorder (33%), conduct disorder (25%),
anxiety disorders (25%), depressive disorders (20%), and learning disabilities (22%). The
comorbidity and recent genomewide scans suggest that ADHD is a polygenic disorders.
The studies for Sibling relative risk and those on twins Show significant genetic influence
inADHD“?

Because ADHD is a common genetic disorder and patients’ behavioral
disabilities can affect not only the person and family, but also the society (school,
workplace, etc.), the influence of ADHD is far-reaching. Also, through the inheritance of
ADHD, similar problems are seen to continue through subsequent generations. For
treatment of ADHD, various approaches, including medications, psychological
remediations, and alternative treatments, have been employed”. Although several
treatments have been successful in ameliorating ADHD symptoms, molecular-based
remedies of ADHD depending on biological explanations are still in primitive stages of

development.

With the increasing growth of high-throughput technology and bioinformatics,
the genetic etiologies of many heritable diseases are unraveling one by one. However,
like many other complex traits, even with high heritability, genetic study of ADHD is at
an early stage. Different from typical family studies of single gene diseases in which
linkages can be detected easily, finding quantitative trait loci for complex multigenic
traits is very difficult even with very dense markers and larger families. Heritable
psychiatric disorders like ADHD are considered one of the most interesting and important
research areas due to possible revelations regarding the genetic background of brain
function, yet there are many difficulties requiring not only profound genetic but also

thorough phenotypic approaches.

Diagnosis and etiology of ADHD

The key characteristic of ADHD is a persistent pattern of inattention and/or
hyperactivity-impulsivity, which are more frequent and severe than behaviors at a
comparable developmental stage. In the case of mental retardation, an additional
diagnosis is made for the child’s mental age. Inattention is also observed in children with
high intelligence when they are placed in academically understimulating environments. If
symptoms are better explained by other mental disorders, ADHD is not diagnosed.
Depending on the Diagnostic and Statistical Manual (DSM-IV), there are three subtypes
of ADHD: the predominantly inattentive subtype, the predominantly hyperactive-
impulsive subtype, and the combined subtype”. The diagnostic criteria for ADHD are

summarized in Table 1.

Table 1. Diagnostic criteria for Attention-Deficit/Hyperactivity Disorderl4
A. Either (I) or (2)
( 1) six (or more) of the following symptoms of inattention have persisted for at least 6
months to a degree that is maladaptive and inconsistent with developmental level:
Inattention
(a) often fails to give close attention to details or makes careless mistakes in
schoolwork, or other activities
(b) often has difficulty sustaining attention in tasks or play activities
(c) often does not seem to listen when spoken to directly
(d) often does not follow through on instructions and fails to finish schoolwork,
chores, or duties in the workplace (not due to oppositional behavior or failure to
understand instructions)
(e) often has difficulty organizing tasks and activities
(f) often avoids, dislikes, or is reluctant to engage in tasks that require sustained
mental effort (such as schoolwork or homework)
(g) often loses things necessary for tasks or activities (e.g., toys, school assignments,
pencils, books, or tools)
(h) is often easily distracted by extraneous stimuli
(i) is often forgetful in daily activities
(2) six (or more) of the following symptoms of hyperactivity-impulsivity have persisted
for at least 6 months to a degree that is maladaptive and inconsistent with developmental
level:
Hyperactivity

(a) often fidgets with hands or feet or squirms in seat

3

(b) often leaves seat in classroom or in other situations in which remaining seated is
expected
(0) often runs about or climbs excessively in situations in which it is inappropriate
(in adolescents or adults, may be limited to subjective feelings or restlessness)

((1) often has difficulty playing or engaging in leisure activities quietly

(e) is often “on the go” or often acts as if “driven by a motor”

(f) often talks excessively
Impulsivity

(g) often blurts out answers before questions have been completed

(h) often has difficulty awaiting turn

(i) often interrupts or intrudes on others (e.g., butts into conversations or games)
B. Some hyperactive-impulsive or inattentive symptoms that caused impairment were
present before age 7 years.
C. Some impairment from the symptom is present in two or more settings (e.g., at school
[or work] and at home).
D. There must be clear evidence of clinically significant impairment in social, academic,
or occupational functioning.
E. The symptoms do not occur exclusively during the course of a Pervasive
Developmental Disorder, Schizophrenia, or other Psychotic Disorder and are not better
accounted for by another mental disorder (e.g., Mood Disorder, Anxiety Disorder,

Dissociative Disorder, or a Personality Disorder).

There are strong genetic factors in the etiology of ADHD. Twin studies suggest

high heritability ranging from 80-88%, and adoption studies using both adopted controls
4

and adopted cases also support the strong genetic component (47% of variance)15 .
Relative risk ratios of ADHD are it E 12-16 for M2 twins, A 3. 5-8 for DZ twins and first-
degree relatives, and A E 2 for second-degree relativesg.

ADHD is a heritable disorder, but environmental factors underlie some causes of
ADHD'é. Those are traumatic brain injury and stroke, severe early deprivation, family

psychosocial adversity, and maternal smoking during pregnancy.

Mode of inheritance and genome wide scans

The mode of inheritance is not clear in complex traits. Usually, polygenic or
multifactorial transmission is suggested. A report describing a segregation analysis of
ADHD rejects the multifactorial polygenic model using likelihood ratio tests”. However,
it is clearly indicated by Morton, N. E., a developer of POINTER, which was used in the
segregation analysis, “Conventional analysis of the mixed model concludes that a major
locus is ‘not proven’, and so the most parsimonious polygenic model may well be
correct.”'8. Because the likelihood does not differ much from each model and a false
major locus model fits almost as well, the likelihood ratio tests may not be appropriate in
this case and polygenic inheritance cannot be rejected in ADHD.

Genome-wide scans also support the polygenic nature of ADHD. The first
genome scan using affected sib pair analysis of 126 pairs in 104 families resulted in no
major gene with highest linkage peak of 2.619. However, their follow up study of 277
affected sib pairs in 203 families found the first major susceptibility locus in a 12 cM
region on chromosome 16p13 with maximum LOD score (MLS) 4.2, p-value = .0000052”.
This result suggests the possibility of major genes in ADHD, but more recent studies of

the group support the polygenic property of ADHD. They found one more susceptibility
5

linkage on l7pll from 270 affected sib pairs in 204 families using 10 cM markers“. In
this report, the linkage signals were MLS of 2.98 for l7pll and MLS of 3.73 for 16pl3
(1 cM markers). With increased samples of 308 affected sib pairs in 226 families, the fine
mapping (~2 cM) of nine susceptibility regions highlighted MLS of 2.55 for 5pl3, MLS
of3.30 for 6q12, 3.73 for 16p13 (same as previous), and MLS of3.63 for 17p1122.

It is notable that the susceptibility regions from this group are completely
different from the genome scans of other groups. A whole-genome scan (~10 cM
markers) in 164 Dutch Sib pairs suggests the linkages in 7pl3 (MLS 3.04), 9q33.3 (MLS
2.05), and 15q15.l (MLS 3.21) using narrow phenotypes”. Also, the genome scan using
a population isolate in Columbia showed significant linkages on 4q13.2, 5q33.3, 11q22,
and l7p11 in individual families“. Taken together, three genome scans suggest different
loci for linkage of ADHD although 17pll is common to two groups.

Depending on the results of current genome scans, there are at least 10 loci or
more that contribute to ADHD. The important basic assumptions of these analyses are; l)
the alleles responsible for ADHD are identical by descent (IBD), 2) there are several
major genes causing ADHD. The first assumption implies that the susceptible alleles are
rare. Moreover, the series of genome scans support the locus heterogeneity of ADHD. If
allelic heterogeneity is also true, then fine mapping narrowing down those regions may
not be possible if not from a single family. The result of fine mapping of l6p13 is
supportive for allelic heterogeneity because finer mapping results in less linkage signals.

These susceptible regions may be partially responsible for ADHD due to family-
specific mutations in the regions. Or, those regions may harbor more causal genes
together than other regions. AS indicated previously, it is not known how many genes are

involved or how they act together. Like other complex traits, it is only clear that a single

6

gene is not responsible for ADHD. Without any knowledge of the mode of inheritance,

the conclusions about causal genes learned only from genome scans can be inappropriate.

Neurobiology of ADHD
The catecholamine system has long been suspected as a main rite of pathology of

25-27 - -
. Studies In

ADHD from neuropharmacology, neuroimaging, and animal models
neuropharmacology were performed on stimulants to increase catecholamine
neurotransmission as well as on non-stimulant”. The stimulant drugs are
dextroamphetamine (d-amphetamine, Dexedrine), methylphenidate (Ritalin), pemoline
(Cylert), and Adderall. Methylphenidate, the most common drug for the treatment of
ADHD, primarily blocks dopamine reuptake but with some releasing effects. The effect
of methylphenidate on norepinephrine is much lesser, and the effect on the serotonin
system is minimal. Non-stimulants include the tricyclic antidepressants, the monoamine
oxidase inhibitors, the aminoketone antidepressants, bupropion (Wellbutrin), the alpha-
adrenergic agonists clonidine (Catapres) and guanfacine (Tenex).

More support for the catecholamine pathology can be found from animal
experiments. Tyrosine hydroxylase gene inactivation results in dopamine deficient mice
that are hypoactivezg, and knock out of the dopamine transporter gene showed high
synaptic dopamine levels causing hyperactivity”. Moreover, animal studies revealed that
selective lesions of the dopaminergic neurons cause significant alteration in attentional
processes3 '.

With neuroimaging studies on nigrostriatal and mesocortical distribution of

dopaminergic neurons in the brain, cognitive impairments in ADHD were suggested due

to a hypodopaminergic state in the prefrontal cortex and hyperdopaminergic state in

7

striatum32'33. One of the clear evidences for the dysfunction of frontO-striatal network in
ADHD pathology comes from an fMRI study using response inhibition tests which are
relevant for ADHD“. They tested response inhibition with and without drug for both
cases and controls. Without drug, frontal activation was greater in ADHD children but
striatal activation was less in ADHD children. However, with drugs, frontal activation
was increased in both groups, but striatal activation was increased in ADHD and reduced
in controls. Taken together, the prefrontal-striatal dysfunction is possibly due to
hyperactive prefrontal region and hypoactive striatum through adrenergic and GABA

system in striatum.

Candidate gene studies of ADHD

Due to the hypothetical pathology as described above, the catecholoamine
pathway has been an important target for previous candidate gene approaches.25‘26'35 Both
dopaminergic and noradrenergic systems are suspected to play roles in modulating the
major psychopathology of ADHD. As Figure 1 shows, there are many candidate genes
that may be responsible for ADHD in the catecholamine pathway.

Research on the dopamine model has focused on genes such as dopamine D2
receptor gene (DRDZ), dopamine transporter gene (DA T1, SLC6A3), dopamine D4
receptor gene (DRD4), dopamine D5 receptor gene (DRD5), and dopamine B-hydroxylase
gene (DBH). Some of the research shows Significant relationships between the specific
alleles of the genes and ADHD, but many of these results could not be replicated in
subsequent studies.35 One of the well-studied genes is SLC 6A3, the dopamine transporter
gene. Most genetic association studies have been done using a variable number of tandem

repeats (VNTR) in the 3’ untranslated region (3’UTR). Many previous studies reported
8

3 -4
6 1However,

that the most common allele, the 10 repeat allele, is associated with ADHD.
a considerable portion of these studies could not find an association between this VNTR
of SLC6A3 and ADHD.38‘42'45 Also, a meta analysis did not reveal significant association
between 10 repeat allele and ADHD.46 As with other examples of association studies
between dopaminergic genes and ADHD, DRD4 and DRDS showed some significant
association although it was not always replicated. Although meta-analyses on DRD4 and
DRD5 found reliable associations with ADHD, further detailed research is needed to

clarify the inconsistency.‘“"47

Catecholamines

Synthesis Delivery Receptors Termination

SynaptosomaI-associated
Protein of25 kDa (SNAP-25)

Ti’rosine Dopamine receptor (DR)
Tyrosine hydro xylase
(TH) 01

L DORA [)2

Do pa decarboxylase
(DOC) Catechol-O-
methyltransferase
D4

Dopamine Bhydroxylase DS
(DBH)

Mon oamine oxidase

' (MAOA a. MAOB)

Dopamine Transporter

(SLC6A3) reuptake

Locus ceruleus

a—Adrenergic synapse (aAR)
Noreplnephnne Va, 021 Ce channels
\ ADRAIA, ADRAIB,ADRAID
PNMT j 052 K channels inhibitory:

Adrenal medulla ADRAZA .ADRMBAURMC
5 -Adrenergic synapse (BAR)
Epinephrine )3 1 K channels inhib'tow' ADRBI
62 K channels inhibitory ADF’B2
Figure l. Catecholamine pathways and candidate genes.

9

The norepinephrine model is also highly favored based on animal models,
pharmacological interventions, and the neural circuitry of attention processes. Among the
genes involved in this process, the (Jig-noradrenergic receptor gene is attractive
particularly because clonidine, an az-noradrenergic receptor (ADRAZA) agonist, is a
treatment drug for ADHD. The biological explanation for this is that the stimulation of
presynaptic otz-noradrenergic receptors results in inhibition of norepinephrine release into
the synapse decreasing hyperactivity and increasing attention span. Association between
one SNP on ADRAZA and ADHD had been examined and shown borderline significance,

but could not be replicated consistently."7"’8'50

Linkage disequilibrium and haplotype studies.

In order to detect minor genetic contributions for mapping complex diseases,
linkage disequilibrium (LD) has been the focus of recent research. With high linkage
disequilibrium, the polymorphisms that are closely linked together form regions that are
called haplotype blocks. Within a haplotype block, several marker alleles in linkage
disequilibrium with the risk allele are enough to map a complex trait. The usual
measurement of LD is D’ or r2, that is significant generally if D’ is higher than .7 and r2 is

51.52
.3.

higher than Large haplotype studies showed that the human genome of the world

population consists of blocks of a few haplotypes with consistent recombinations.53'SS
However, a simulation study suggests genetic drift may generate block-like patterns of

linkage disequilibrium.56 Also, a linkage disequilibrium map of chromosome 22 revealed

that many susceptible gene regions of schizophrenia did not Show high LD.S7 These

10

results suggest that high density maps of disease loci are needed for mapping complex
traits, such as ADHD.

Many haplotype studies of complex traits are ongoing and most employ a set of
several haplotype-tag SNPS for testing association between haplotypes and diseases.
Generally, haplotype approaches show better associations than approaches using single
marker polymorphisms. As a target of methylphenidate, the stimulant drug for treatment
of ADHD, SLC 6A3, the dopamine transporter gene, is studied frequently.

SLC6A3, located in 5p15.33 telomeric region, spans 52,500 bps with 15 exons.
The region from 10,000 bp upstream to 2,000 bp downstream contains a total of 337
SNPS, although most of them may be sequencing errors or rare mutations. A relatively
high-density linkage disequilibrium map was constructed over this gene region and
shows two clear blocks within the gene.58 The second block beginning before exon 9 can
be subdivided into two more blocks. Their haplotype association study on bipolar
disorder revealed better association results than single marker or several markers located
close to each other.59 The first (5’) haplotype block of SLC6A3 did not show any
significant association, but the second (3’) block showed some significant association at a
p-value < .05 level through the entire block with most of the SNP combinations. However,
with most SNPS typed in the whole second LD block, the haplotype showed the most
significant association by extended transmission disequilibrium test (ETDT). Haplotype
association studies between ADHD and the SLC6A3 gene have also shown better
association results. A haplotype consisting of three polymorphisms, exon 9 SNP, intron 9
SNP, and the 3’UTR VNTR, was associated significantly with ADHD using the
transmission disequilibrium test (TDT)37. Another haplotype study of a larger region of

SLC6A 3 revealed a significantly biased transmission of a haplotype“).
I l

This better association using haplotypes was hypothesized to result from more
sensitive detection due to the higher possibility of capturing the disease allele within a
haplotype than a single marker. However, if the haplotype results were looked at more
closely, sometimes the haplotype association studies showed significance even though a

(’0 Also, a set of

set of polymorphisms in low linkage disequilibrium was used.
polymorphisms that are in high linkage disequilibrium did not show a higher significance
than using single marker allele.59 It seems that some of the significant haplotype results
might come from the combined effect of two or more different disease polymorphisms in
different haplotype blocks. It should also be noted that the Significance of association
results could be strikingly different depending on which set of polymorphisms is chosen
even in the same linkage disequilibrium blocksg‘sg.

Some other studies suggested other possible effective polymorphisms.("’(’3 It is
reasonable to think that there are several polymorphisms in the SLC 6A3 gene locus that
may act on expression, stability or other effects. A gene expression study using
haplotypes of SLC6A3 showed that promoter and intronic variants affect the
transcriptional regulation of SLC6A3 and suggested that particular combinations of
polymorphisms in haplotypes affect the expression."4 These results suggest that more

careful approaches are needed in haplotype association studies considering not only the

block size and LD but also the number of effective polymorphisms.

Phenotypic considerations.
With heritable complex traits of unknown etiology, the exact phenotype
characterization is an important issue for genetic studies. Because there is no

demonstration of consistent neurobiological differences in ADHD children, the

12

controversial phenotype definition and etiological heterogeneity may be the reason of
invalidity in the genetic study. For ADHD, the fourth edition of diagnostic and statistical
manual of mental disorders (DSM-IV) defines ADHD phenotypes as three subtypes,
ADHD-combined type (ADHD-C), ADHD-predominantly inattentive type (ADHD-PI),
and ADHD-predominantly hyperactive-impulsive type”. Research to date has been done
on ADHD-C because it is the most prevalent. The ADHD-C and ADHD-PI are different
cognitively and in familial historyés'm’, although those are not differentiated consistently

(’7‘68. These subtypes are coded on the basis of two different

by the neuropsycological data
symptoms, inattention and hyperactivity-impulsivity that can be also considered possible
separate phenotypes.

The consideration of phenotypes leads to the necessity of finding a consistent
measurement for the genetic approach. One of the notable approaches is endophenotype.
This concept came from the genetic theory of schizophrenia, having the synonymous
meanings as “intermediate phenotype”, “biological marker”, and “subclinical

”69. It can be defined as etiologically pure phenotype correlated with ADHD

phenotype
symptoms that is familial and appears in unaffected relatives.70 The endophenotype
should be associated with candidate genes and heritable. Because the biological
endophenotypes are relatively more expensive to measure than congnitive
endophenotypes, cognitive endophenotypes can be considered first.

Among several putative endophenotypes suggested, it is notable that the
dysfunction of the response inhibition may be one of main etiologies in ADHD.”"7"72
Disinhibition can be conceptualized as fast but inaccurate response, response

perseveration, and a failure to respond appropriately in a response conflict task.73 The

possible endophenotypes of disinhibition in ADHD children are varied and were tested

13

for the possibility of a familial neuropsychological endophenotype.“ Although those
results show promising cognitive endophenotypes, it would explain the etiology of
ADHD more precisely if the neurobiological function of the endophenotypes could be
investigated in depth using neuroimaging or neurophysiological measurement.

One interesting feature of ADHD is the difference in prevalence between girls
and boys. The ratio of boys to girls ranges from 3:1 to 8:1, and the ratio is higher in cases
of clinically referred ADHD. Meta analysis on the gender difference in ADHD found that
ADHD girls showed lower hyperactivity, fewer conduct disorders, lower externalizing
behavior, and greater intellectual impairment (restricted to clinic-referred children), but
there was no gender difference in impulsivity, academic performance, social functioning,
and fine motor skills although most data were limited only to clinic-referred samplesm".

There are several hypotheses to explain the greater occurrence of boys with
general childhood psychopathology. The most probable ones are the polygenic multiple
threshold model and constitutional variability model. The former explains that girls need
more genetic risk factors to be affected than boys and the latter describes that different
casual factors affect females and males differently. The statistical test for the two models
reveals an inclination to the polygenic multiple threshold model”. However, the
difference in cognitive function between girls and boys suggests the possibility of the
constitutional variability model. Moreover, recent research to find quantitative trait loci
(QTL) related to cardiovascular functions using consomic rats showed that considerably
different loci were related to cardiovascular function between women and men supporting

176'77. Study of genetic contributions

the possibility of the constitutional variability mode
to diseases which show gender specific predisposition may need to examine whether

some risk alleles are gender specific.

14

I

gen‘er
and dc'
Cami:
C0326)
oxacc.
liit til)“
arms .1
dopznr

mhcr..

 

Caitlittt ,;
eiaCiliah
gender.
gonadct’
mannuut

ICiLlit‘d j,

FArp v! ‘ .
~

 

As a strong candidate gene of ADHD, the dopamine transporter varies across
genders in respect to the expression and density of protein. In a rat study, the mRNA level
and density of the dopamine transporter was significantly higher in females than males.78
Combined with previous studies of the same group, it is suggested that such difference
comes from a genomic effect of female gonadal steroids by comparison between

. . . 7 , (
ovarIectomIzed females and Intact males. 98’

More interestingly, mRNA expression of
the dopamine transporter is not regulated by estrogen in several brain regions including
some striatum regions of female rats.8| Ovariectomy in adult female rats reduces the
dopamine transporter density but increases mRNA level, suggesting the involvement of
other cellular mechanisms.82 Also, in a human study, SPECT results Show significantly
higher density of the dopamine transporter in the striatal region of females.83

Gender difference in adrenergic receptors has been reported through
cardiovascular studies.84 Studies, that Show antagonists for ctz-adrenoceptor affect male
ejaculatory function, suggest further differences of the adrenergic system between
genders.85 Another study using an antagonist for ag-adrenoceptor for tail artery of
gonadectomy rats showed that gender differences in ong-adrenoceptor function are not
maintained by gonadal steroid hormones suggesting that the gender difference may be
developmentally regulated.“

Although it has not been focused well, for association studies of candidate genes

related to the catecholamine pathway, gender is a very important factor to consider in

regards to ADHD.

15

 

 

Stud)

”31...”.

CULT:

mutt}.

r)
{.2
:3
CA.

.lliR. :

\\ CIL’ .

 

Study Design

Although ADHD is highly heritable, genome-wide scans did not find a strong
linkage nor any replicated regions that appeared interestingm‘z". Inconsistent results are
seen frequently in candidate gene approaches. The reason for the series of inconsistent
results seems to come from the lack of a major gene effect reinforcing the multigenic
nature of ADHD, which implies minor contributions from a large number of genes.
Previous research indicates that the genetic study of ADHD requires more elaborate
methods to determine genetic etiology.

Without clear pathology of ADHD, genome scans are attractive. However,
currently, there is no good method for finding relevant genes in complex traits that are
mutifactorial with heterogeneity. From the precept of previous genome scans of ADHD,
candidate gene approaches were tried instead. For candidate genes, two drug target genes
ADRAZA and SLC6A3, as well as DRD4, which has shown the most reliable association,

were selected for candidate genes.

16

Inirm
for A:
locatt
trans.
genC'
lip dl~
exon.
llISI. .

Ilaplii-

 

Candn

ADHD
IEQIOn.

Proms;

 

Chapter 2

Candidate genes and initial analysis

Introduction

As described in the previous chapter, three candidate genes selected are relevant
for ADHD, so these have been the focus of candidate gene studies of ADHD. ADRAZA,
located in 10q24-26 in the middle of the chromosome, consists of one exon with a
transcript of 3650 bp. SLC6A 3 is located in 5p15.3 near the end of the chromosome. This
gene is quite large, approximately 52,640 bp consisting of fifteen exons. DRD4 is 3398
bp also located in the telomeric region of chromosome 11 (11p15.5), and consists of four
exons. In this chapter, the current information about the candidate genes is summarized
first, and the analyzed data for the polymorphisms selected initially are described.

Haplotype analysis was done on ADRA 2A and is discussed in the later part of this chapter.

Candidate genes
a-ZA adrenergic receptor gene (ADRA 2A)

In molecular genetic approaches to ADHD, the most obvious target has been the
catecholamine pathway, in part because it is the site of action of psychostimulants used to
treat ADHD.87 As a result, both dopaminergic (DA) and noradrenergic (NE) systems,
which modulate one another, are thought to play roles in shaping the pathophysiology of
ADHD. Both systems are expressed in the prefrontal cortex and its many projection
regions. Accordingly, a number of prior studies have investigated DA genes with
promising, but small, effects for DRD4, DRD5 and SLC6A34"~“~38.

In contrast, relatively little research has examined NE-relevant genes.

17

 

Ire

 

Xx lit.
lilL’ it‘
non:
CHI:
I
ADI 5 .

alph;

 

Noradrenergic neurons in the brain are concentrated in the brain-stem nucleus known as
the locus coeruleus (LC). They project throughout the brain, providing the only source of
noradrenergic stimulation to the prefrontal cortex and thus key NE-relevant genes are
expressed in the prefrontal cortex and other brain regions relevant to the development of
ADHD. Three types of noradrenergic receptors are traditionally recognized, alpha-l,
alpha-2, and beta. Animal research suggests that NE projections in the prefrontal cortex
enhance prefrontal cortical function primarily through post-synaptic alpha-2 receptors”.
Of the several types of alpha-2 receptors in the brain, the most promising candidate for
study is the ot-ZA adrenergic receptor (ADRAZA). This receptor is expressed in many
areas of the brain, but is the most prevalent NE receptor type in the prefrontal cortex.

It is now relatively well established that NE is important to functions of the
prefrontal cortex that are implicated as core deficits associated with ADHD, including
working memory, focused attention, and response controlgo. As noted by Berridgegl,
substantial data suggest that NE neurons are important in the regulation of arousal,
wakefulness, and signal-to-noise ratio in attention. NE thus supports a key vigilance
system in the brain”. The importance of NE to vigilance, alertness, and state regulation
suggests its involvement in ADHD because difficulty with arousal and activation are core

features of several theories of ADHD”'95

and are noted as needing explanation in other
theories“). As a result, dysfunction of the ascending NE system has often been theorized
to mediate ADHD97'98‘99"00. These theories are supported by substantial behavioral
evidence suggesting that deficits in arousal and alertness are linked to ADHD. This
evidence includes excess slow wave activity on EEGS'O', evidence of impaired signal

. . . . 02
detection usrng the d-prime parameter on Continuous Performance Testsl' , and slow and

variable reaction times on fast reaction time tests in children with ADHDWOMO". All of

18

13$k>
C>?C..
aflHUJ
prefro
used \
recept
Shontt
PICI‘TUI

behuxi

 

.Ililill
Undern

IfCEpjn

gene a:

I3lhluj5‘

these findings are consistent with abnormal functioning of a vigilance/arousal system that
is likely mediated by ascending NE neurons, of which ADRA2A plays a key role in the
prefrontal cortex.

Recent work implicates NE (as well as dopamine), and ADRAZA in particular, in
tasks that reflect executive functioning in animals89 and humans'“. These functions,
especially working memory, are involved in ADHD96. Pharmacological evidence in
animals and humans also supports the role of NE, and in particular ADRAZA, in the
prefrontal cortex and thus potentially in ADHD. The 0t-2A agonist clonidine has been
used widely in the treatment of ADHD childrenm”, suggesting a potential role for the
receptor in symptom expression. More definitive evidence emerges from recent work
showing that the selective ot-2A agonist guanfacine improves function on tasks reliant on

90.l07

prefrontal cortical functions in monkeys and in humans'o", but does not affect

9f)

behavior when the prefrontal cortex is not challenged . Thus, pharrnacologic

investigations point to an important role for the NE system, especially the ot-2A receptor,
in the cognitive operations of the prefrontal cortex that are suspected of involvement in
ADHD.

In short, there is ample evidence to suggest that NE neurons are important in

'08, and at this initial stage of

ADHD and its associated multiple cognitive deficits
understanding an important NE receptor in the prefrontal cortex appears to be the ot-2A
receptor. It is therefore important to evaluate whether polymorphisms of the ADRAZA
gene are related to ADHD in order to set the stage for further etiological studies.

Although the investigation of ADRA2A has only begun in relation to ADHD,

association between ADHD or its symptoms and one SNP in the ADRAZA gene,

r31800544 (which creates an Mspl restriction fragment length polymorphism (RFLP)),
19

nOIL'iti

fiOEI'
5}':I:;i
poi} T‘.‘
ltnfii

Alllf
ilHlR
“‘3: .9
all Cit;
laltcm
10 dm

due to

pit“ ltlt

Dflptllll

Effie, _\
IVXIR
most Cl
Considc
510w

bel“ cc”

 

 

has been examined in four published studies. Comings, et al.48 examined this association
in children with Tourette’s Syndrome, and found that the additive score of three
noradrenergic genes correlated with expression of ADHD symptoms. A follow up report
from this sample found that allele m of this SNP in ADRA2A was associated with ADHD
symptoms”. However, Xu, et al. failed to find linkage and association with the same
polymorphism using a transmission disequilibrium test (TDT) analysis in 94 nuclear
families in which the proband had ADHDIOQ. Roman, et al. studied 96 children with
ADHD and their parents in a sample from Brazil. Although their haplotype relative risk
(HRR) analysis with the disorder also yielded non-significant effects, this polymorphism
was associated with ratings of inattention and hyperactivity, suggesting the possibility of
an effect of the gene on symptom expression”. These two results both evaluated the G/G
(alternatively denoted as m/m) genotype as the risk genotype. Nonetheless, it is difficult
to draw clear conclusions about ADHD and ADRAZA from these few preliminary studies
due to conflicting findings and the fact that a sample of Tourette’s Syndrome patients

provide the main positive findings, which may not generalize to other ADHD samples.

Dopamine transporter gene (SI. C 6A 3)

One well-studied candidate gene on ADHD is SLC 6A3, the dopamine transporter
gene. Most genetic association studies have used a variable number of tandem repeat
(VNTR) on the 3’ untranslated region (3’UTR). Many previous studies reported that the

3(-
’4’. However, a

mOSt common allele, the 10 repeat allele, is associated with ADHD
COnSiderable portion of these studies could not find an association between this VNTR of
S[(36143 and ADHD38‘42'45. Also, a meta-analysis did not reveal significant association

hmWeen the 10 repeat allele and ADHD“. However, neuroimaging studies suggest the

20

 

mi...

famil)

dOptn

 

dUPdfl
IIS m
Width
dOpar
differ

brunt

Sequm

involvement of the dopamine transporter (DAT) in the major etiology of ADHD'”). A
study using single photon emission computed tomography (SPECT) showed that ADHD
patients (four women and two men) have an increase of 70% in dopamine transporter
density over controls (total 30) with age-correction in striatum, suggesting the

””12. It is also known

involvement of the dopamine transporter in the etiology of ADHD
that the therapeutic treatment with methylphenidate reduces the increased DAT
availability in ADHD adult patients‘ '3.

The dopamine transporter is a member of a Na+w and Cl' —dependent transporter
family, in forms of disulfide-linked homooligomer in membranes. It is known that the
dopamine transporter interacts with the protein kinase C-alpha binding domain. There is
direct evidence of phosphorylation and its regulation by PKC and MAP kinase, and it has
several sites for N-linked glycosylation in the large second extracellular loop'”. The
dopamine transporter is located in the synaptic craft and highly expressed in the midbrain.
Its main role is modulation of dopaminergic neurotransmission by the reuptake of
released dopamine. It is supposed that some RNA editing occurs in the brain, and the
dopamine transporter has several relevant protein sequences. Ensemble predicted three
different mRNAS, but a study using rats could not find any alternative splicing in some
brain regions”. The 12 transmembrane domains were well predicted from the multiple
sequence alignment of the related transporters'm‘m. The strongly preferred
transmembrane prediction suggests the 12 transmembrane domains starting from inside to

outside of N to C terminal, and the positions of the transmembrane region in each

sequence are identical (Figure 2).

21

 

 

'I‘I’X‘T x-o
-' r-r r ‘1 -. Threr-i;
Q

5‘

billt' i\ ‘

 

it till 1*

block

DUPLH

bt’llim
bl (i,
di’l‘dm

Willem

Extracellular

090000000

 

eeeeE
«as;
eeeeee

o
éfifig

 

 

 

 

' 0

oeeooo " 3' ° - $2 3
two We 6:9 m 090 mi) @5239 woo @553 moo cm
W Give em Q99 063 We com me @719 09o W
cm m cm We ow (”can @659 mm 699 Game @559
W an» owe me me» We new we we @c’to com
e09 We ow 0% ow mm W Ono w more we
cow 0% @3610 Gina W cm cow 613» com me @639
. _ 0 0000000! ., s G 0 o
-- '- W Li
" ‘ “3,23%? 3 «3

30000003

   

Cgtoplasn
Figure 2. Schematic representation of dopamine transporter

(http://pharmacogenetics.ucsf.edu/set l /DAT/)

As described in chapter 1, the rather large SLC 6A3 gene region contains two LD
blocks”. Depending on the literature, the second block is associated with bipolar disorder
with p-value less than .05. Haplotypes consisting with three polymorphisms in the second
block also result in Significant association with ADHD through TDT. However, as
mentioned in chapter 1, it is not clear if there is a functional polymorphism residing in the
haplotype or the combined effect of several functional polymorphisms causes the

association.

Dopamine receptor D4 gene (DRD4)

Dopamine receptor D4 is one of five subtypes of dopamine receptors, which
belongs to the G-protein coupled receptor 1 family. The action of this protein is mediated
by G proteins that inhibit adenylyl cyclase. Like other subtypes of dopamine receptors,
dOpamine receptor D4 contains seven putative transmembrane domains. However, this

protein contains repeat variants that change the length of the protein in the putative

22

C\.'i{ T

\ari.

alter:

‘ldfll
V \.

Then

four : *
[w p.
sumrtx
alléih >2
\NTR
discmt
Chlide‘i
repeal 3
of \'\T

0n the c

associati
hapionp
ShOu‘ed \

on {his L'.

 

cytoplasmic part after the last putative membrane domain. In the genomic region, this
variant is located in exon 3 and is the focus of association studies of psychiatric disorders.
Brain tissue examination showed that this gene is not imprinted in the human

- us
brain .

However, unlike the dopamine transporter gene that does not have any
alternatively spliced isoforms in superior cervical sympathetic ganglia and dorsal root
ganglia, alternative splicing transcripts for DRD4 were found in dorsal root ganglia”.
There is no transcript of DRD4 in cervical sympathetic ganglia.

Dopamine receptor D4 gene is located in llp15.5 near the telomere. There are
four exons on the gene, spaced over 3398 bp. Most association studies have been done on
two polymorphisms, 120 bp repeat promoter polymorphism and exon3 VNTR. As well
summarized in a review article”), the association results are inconsistent although meta-

46'47. interestingly, there is a report that, within

analysis showed the association of DRD4
VNTR subtypes, over 10 percent of ADHD probands have rare subtypes that were not
discovered in the previous population studies'zo. Also, in Chinese Han population, ADHD
Children with normal IQ and methylphenidate responders showed the association of 2
repeat allele using ethnically matched controlsm. These suggest that allelic heterogeneity
of VNTR may contribute to the association and the subtypes may be different depending
on the ethnicity.

TDTs using several more polymorphisms in the promoter region showed an
association between —616 SNP and ADHD with p-value of .008, rather than no
association of 120 bp insertion/deletion (l/D) polymorphism or VNTR'ZZ. There is no
haplotype association study yet for DRD4. However, the LD structure of this gene region
showed strong LD among the 7 repeat allele of VNTR for evidence of positive selection

23

 

he

 

ADHD:

Sam;

mcx
gm:
mew
Obi.
n=h
llt‘dit
D\A

Pmbri
mmm-
mmmg
Ihelfld.
hth
(mamr

Efihm

BChaH'“

Rating \
\

 

Sample collection and demographic description

DNA samples were requested from affected children, their biological parents and
one sibling nearest in age when possible. A total of 177 probands were studied in three
groups: Non-ADHD Control (n=62), ADHD-C (n=81), and ADHD-PI (n=34). The
majority of the probands were Caucasians (82%). The children were aged 7—13 (mean =
9.6), and included both boys and girls (64.5% boys). Complete trios were obtained for
n=107 families. For buccal DNA preparation, a modified method described by
Meulenbelt was performed in which cheek swabs were used for sampling followed by
DNA preparation using phenol/chloroform purification (average 60 ug DNA per
collection) 125.

A regular multistage recruitment and screening procedure was used to identify
probands, based on the methods of the MTA studies. Families were recruited from the
community using public advertisements and mailings to all parents of children in 2nd
through 6th grades in the local school district. They were ruled out from participating if
the index child had autistic disorder, bipolar disorder, Tourette’s Syndrome, psychosis,
history of head injury with loss of consciousness, history of seizures, or full scale IQ < 75
(evaluated with a 4-subtest short form of the Wechsler Intelligence Scale for Children, 3rd
Edition).126

Index children were considered as possible ADHD if they either passed prescreen
cut-offs on both parent and teacher versions of common ADHD rating instruments (Child
Behavior Checklist or Teacher Report Form, Behavior Assessment Scale for Children

127.128

Rating Scale, or DSM-IV symptoms checklist)‘29 or were previously diagnosed as

ADHD (any type) by a physician or psychologist in the community who utilized teacher
24

mdj
COEZT

UC\C'

com?
an "(I
mix
PCiCCI
One:

Hort:

 

arms
S}‘np:
e\ClUi

mane

lniiia

do

”J
L)
:3

Ch0\er

and parent ratings to arrive at their diagnosis. Children were considered as possible
controls if they were below cut offs on all of these parent and teacher scales and were
never diagnosed with ADHD.

Final diagnosis was then determined after administration to the primary caregiver
(usually the mother) of a structured diagnostic interview, the NIMH Diagnostic Interview
Schedule (DISC-IV) for DSM-IV."7 The DISC-IV is a widely used and accepted
instrument with acceptable reliability and validity for evaluating diagnoses in community
samples. Inter-rater reliability for ADHD diagnosis in our study was k=l.0, due to the
computer-assisted nature of the interview procedure. After administration of the DISC-IV,
an “or” algorithm was employed to identify ADHD.I30 If the child met onset, duration,
and impairment criteria, had at least 4 symptoms on the DISC-IV, and exceeded the 90th
percentile on teacher cut-offs, then a symptom was counted as present if it was endorsed
on either the DISC-IV by the parent or by the teacher on the DSM-IV checklist
.(“sometimes” or “often” rated as “present”). In that way, a final symptom count was
arrived at for each child and they were assigned to either the control group (4 or fewer
symptoms) or one of the ADHD groups. Children with ADHD-hyperactive type were
excluded as explained earlier. Also excluded were children with 5 symptoms of either

inattention or hyperactivity, because their subtype status is indeterminate.I30

Initial association results

For the preliminary association study of the dopamine transporter gene and
dOpamine receptor D4 gene, the most extensively investigated polymorphisms were
chOSen. For the non-stimulant medication system of ADHD on the noradrenergic system,

Mspl RFLP, which has been studied mostly on a-2A-adrenergic receptor gene, was
25

Chi“;

and

CWT;

 

 

ADII
r013"
ADI:

5

[WIT

21550:

S.\‘P_~

 

Tabic

(it

DR

S]. (

m

chosen first. For this SNP, Comings et al.48 examined children with Tourette’s Syndrome,
and found that the additive score of three noradrenergic genes was correlated with
expression of ADHD symptoms. A follow up reported that allele m was associated with
ADHD symptoms.49 However, Xu et al. failed to find linkage with the same
polymorphism using a TDT analysis in 94 nuclear families in which the proband had
ADHDW). Roman et al. studied 96 children with ADHD and their parents in a sample
from Brazil. Their HRR analysis also yielded non-significant effects, however the risk
polymorphism was associated with ratings of inattention and hyperactivity, suggesting
the possibility of a weak effect of the gene on symptom expressions”. Those two results
both demonstrated that the 0/6 (denoted as m/m) genotype is the risk genotype. In a case
of small gene effect on ADHD, one SNP as a marker cannot show a significant
association. So, all the publicly reported SNPS on the region were screened and two more

SNPS were selected for this association study.

Table 2. Genotype Association Results (p-values of chi-square test).

 

 

 

 

_ Control vs Control vs
Gene Polymorphism
ADHD-C + ADHD-PI ADHD-C
Insertion/deletion .01 .005
VNTR .90 .97
Exon 9 .24 .24
SLC6A3 Intron9 _ ., .. . ,, . .07 _ , . , .079 . .. .. ..
\ rsl 800544 (Mspl) .77 .65
ADRAZA ’ rg‘1‘360‘5‘157ma) ' " ' .83 f if .67
yrs5'5396i68m(iDrril)‘ .. .85 .. . . , .77

 

 

26

 

inst“

li0l .;

 

that 7

$50.

or tl‘.

 

l‘0l 3~ n

Sln ll

-.-_1
\

Our genetic data revealed strong association with DRD4 insertion (Table 2). The
insertion/deletion polymorphism on the DRD4 promoter region has been studied before
for association with ADHD and the insertion allele is significantly associated with
ADHD'“. Our association result replicated this. We found that the SLC6A3 VNTR 9
repeats associated with ADHD in this study samples (Table 2), however most association
studies of the VNTR in SLC6A3 showed 10 repeats as the risk allele. In our study, all
alleles other than 10 repeats were significantly associated with ADHD. Another group
that found the same result with this VNTR“, and meta-analysis did not find a significant
association between 10 repeats and ADHD“. This inconsistent result is not limited to
ADHD. Other psychiatric or neuroscience research also found inconsistent association
results for the VNTR of SLC6A3‘32. It can be hypothesized either that there is a real
acting polymorphism which has a different pattern in linkage disequilibrium with VNTR
or that the different VNTR subtypes make the difference working as an effective
polymorphism.

Interestingly, the transmission disequilibrium test (TDT) did not show any
significance between ADHD and either DRD4 or SLC 6A3. An association trial on parent
groups found more significance between the parents of controls and the parents of
probands. For the DRD4 insertion/deletion polymorphism, the p-value was .007, andfor
SLC6A3 VNTR polymorphism the p—value was .03. This means that the parent groups are
already sorted significantly in the risk polymorphisms. With consideration of high
heritability, it is reasonable that the parent group of probands has more risk genes because
they may have expressed ADHD as children.

Despite the DSM-IV’s identification of ADHD as a categorical disorder, many

genetic analyses indicate that ADHD symptom dimensions have the same genetic

27

 

chm:

 

ildhiL

lih(ll

Did

-l 1’.)J

influences at all levels of severitym.

Those data commend consideration of the
association of genetic markers with dimensional symptom ratings. We conducted such
analyses using the quantitative transmission disequilibrium test (QTDT), with the
advantage of considerably greater statistical power than is available in the conventional
TDT analysis and inclusion of parent symptom counts. This approach revealed a trend
toward significance using the Mspl polymorphism (p=.022) and confirmed linkage
between the T allele of DraI and symptoms of inattention (p=.003). The T allele of Dral
is found primarily on a subset of chromosomes containing the G allele of the Mspl

polymorphism. The results of QTDT suggest that the risk allele might reside on the

chromosomes containing both the T allele of DraI and the G allele of Mspl.

Table 3. TDT results for each ADRA2A SNP and ADHD subtype.

 

 

 

 

SNPS ADHD type T NT RR X2 P value

ADHD-C l4 7 2.00 2.33 .l3

Mspl (G allele) ADHD-PI 8 5 1.60 0.69 .41
ADHD-(C-l-PI) 22 l2 1.83 2.94 .086

ADHD-C 6 3 2.00 1.00 .32

Hhal (G allele) ADHD-PI 2 3 0.67 0.20 .65
ADHD-(C+PI) 8 6 1.33 0.29 .59
ADHD-C ll 3 3.67 4.57 .033

DraI (T allele) ADHD-PI 7 4 1.75 0.82 .37
ADHD-(C-l-PI) l8 7 2.57 4.84 .028

 

 

ADHD-C: combined type. ADHD-Pl: primarily inattentive. ADHD-(C+PI): both ofthe types. T:

transmitted. NT: non-transmitted. RR: relative risk.

Using TDT, we found an association between ADHD and Dral RFLP of the

ADRA 2A gene that did not show any significance in the case-control association study
28

 

ill?

"i v‘
4 \
ll“... .

less '

Yr-
LD..

 

‘th .

(Table 3). The less common allele of the Dral polymorphism was preferentially
transmitted to ADHD children. It is interesting that this preferentially transmitted allele is
less frequent in the athlete endurance groupm. This result suggests that the DraI RFLP
may contribute only a minor portion to ADHD and could not be sorted in the case-control
study, but instead preferential transmission was seen from the even distribution of parent

groups.

Table 4. QTDT results for ADHD symptom dimensions

 

 

 

 

 

 

 

ADHD type Marker Allele* X2(df) P
Mspl G 5.33(1) .022

Inattention Hhal A 0.00( 1) NS
DraI T 910(1) .003
Mspl G 485(1) .037

Hyperactivity-impulsivity Hhal A 0.05(1) NS
DraI T 6.95( 1) .015

 

Empirical p values are presented. NS: not significant.

*The allele conferring increased risk is denoted.

These results suggest that the differentiated effect between the major contribution
and the minor contribution of genes should be considered. In case of the relatively major
contributing polymorphisms, the case-control study might be useful, and, for detecting
the minor contribution of polymorphisms, the transmission disequilibrium test would be
helpful. The major and minor contributions can be different depending on the population
sample, as an example, the clinic-referred children and the community-recruited children
need to be considered differently. One more important point is that there needs to be a

consideration of the moderate contribution of some polymorphisms that cannot be sorted

29

 

600..

TDT

Initi;

 

to 3th
in 11:.
being

Q

tuncti

 

the ca

enough between parent groups and cannot be distributed evenly enough to be detected by

TDT.

Initial haplotype analysis on ADRAZA

The variable results from previous association studies on ADRAZA may also be
due to the difficulty of detecting minor contributions of a particular candidate gene to the
liability for developing ADHD. This limitation is accentuated by restricting the analysis
to the examination of a single polymorphism in the gene. If the selected polymorphism is,
in fact, the only or primary functional polymorphism that contributes to the disorder
being studied, it will then be the most robust marker for the disease. If it is not the
functional polymorphism, however, then it would serve merely as a surrogate marker for
the causative allele, and yield less robust findings in studies of association and linkage.

No evidence suggests that the rsl800544 SNP is functional. Therefore, to address
this concern, we chose to examine more closely the haplotype structure of multiple
markers in the ADRAZA gene and identify a set of SNPS to study. Previously, our case-
control association studies on ADRAZA did not provide any significant results. Therefore,
we used the transmission disequilibrium test (TDT) and quantitative transmission
disequilibrium test (QTDT) to assess association and linkage of the ADRA2A gene
polymorphisms with ADHD in two of its subtypes (ADHD-C and ADHD-PI) and in its

two core symptom dimensions (inattentiveness and hyperactivity-impulsivity).

SNP selection
ADRAZA is a small gene with a genomic size of <4000 bp. The SNP Consortium

(TSC) database identifies 12 variants, 8 of which are within or near the mRNA genomic

30

 

 

rslSH
pncn
in 0b

P0l}n

 

chlc
ll§i.\l

RFLP

three
Equfli
aPPCdl

“gnu?

region. A summary of the SNPS that have literature reports or frequency information is
provided in Table 5. We include in Table 5 the frequency information from our own
sample in the current study as well. The Mspl RFLP is located 5’ of the transcribed region
and the allele frequencies are similar in all Caucasian groups reported. There is a
polymorphic SNP in the 5’-UTR, 3 non-synonymous mutations in the coding region, and
two 3’UTR SNPS. As can be seen in Table 5, the allele frequencies observed in the study
sample (labeled “Michigan” in the table) were typical of those reported in the literature.

We first examined the polymorphic status of the non-synonymous, coding SNPS,
rsl800034, r51800035 and r5180036, because variants at these positions have the
potential to produce functional differences in the protein. We did not find these variants
in our population, reinforcing the suspicion that these may be rare mutations. Three
polymorphisms were chosen for analysis of the association with ADHD, based on their
allele frequencies and their spacing in the genomic region. These were the Mspl RF LP
(rsl800544) previously studied, 3 leal RFLP (r51800545) in the 5’ UTR. and a Dral
RFLP (r5553668) in the 3’ UTR ofthe ADRA2A mRNA (Figure 3b).

As previously described, participants in 177 families were genotyped for the
three SNPS in ADRAZA. For each of the markers, we evaluated Hardy-Weinberg
equilibrium by simulation, using 10,000 iterations for each simulation. All of the markers
appeared to be in Hardy-Weinberg equilibrium, as their one-tailed p-values were all non-

Significant (i.e., Mspl p = .355, Dral p = .343, Hhal p =.7l9).

31

Table 5. Polymorphism information for the ADRA2A locus.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SNPs Frequencies Population Reference

C (M allele) G (m allele)
$18005“ .26 Caucasian 5” (Roman et al. 2003)
5 promoter .71 .29 Caucasian '35 (Lario et al. 1997)
gig)“ I .74 .26 Canadian “)0 (Xu et al. 2001)
RFLP .73 .27 French, Irish & Scot Canvas Database

.67 .33 Mostly Caucasian Michigan: Controls
rsl800545 G A
5’ UTR .89 .l 1 French & Irish Canvas Database
Hhal RFLP .88 .12 Mostly Caucasian Michigan: Controls
rs 1 800034 Mutation '3" (Feng et al. 1998)
rsl800035 Mutation "i" (Feng et al. 1998)
rsl 800038 C A
synonymou .72 .28 Random '36 (Feng et al. 1998)
5 change .71 .29 Japanese JSNP Database
rsl800036 Mutation '3" (Feng et al. 1998)
Dra 1 RF LP C T
r8553668 .l9 Caucasian I37(Hoehe et al. 1988)
3, UTR .20 Caucasian I34(Wolfarth et al. 2000)

.81 .19 Mostly Caucasian Michigan: Controls
rs3750625 C A
3 ’ UTR .73 .27 Japanese JSNP Database

32

 

 

Mk
.8 it lHl
ll . .l. :10 luw
Ill .. H .... in .0 1.1m
. . . H. till i\ a .
v . .Nttu flu . t\|m PL 1)
t1 l . .
0nd .18... 10.
a»!

1

than 1

Linkage disequilibrium and haplotypes

The linkage disequilibrium and haplotype were studied on three SNPS of the
ADRAZA region. Linkage disequilibrium (LD) was calculated using D’ and r2 for each
pairwise combination of SNPS (as shown in Figure la). Calculated from the GOLD
software package“, both D’ and r2 values show significant linkage disequilibrium
between Mspl RFLP and Dral RFLP, and it is likely that the Mspl RFLP shows a
significant association in QTDT because this site is in linkage disequilibrium with the
Dral RFLP (Figure 3a). The D’ value between the Hhal RF LP and Dral RFLP is highest,
while r2 is very low. On the other hand, although the D’ value between the Hhal RFLP
and the Dral RFLP was high, the r2 was very low. This is thought to occur when the rare
allele at one locus is linked to the common allele at the other locus and vice versa, rather

than linkage occurring between alleles of similar frequency. '39

33

a.\

'73"

 

 

C

.()h

light
Panel

b \h()\
llnt‘ (11‘.
9070:3144.

Hailit'tt

a. Marker linkage

 

 

D’ r2
Mspl Hhal Dral
I 1035 ”P I .827 .196
' 3°76 b” ' .793 .351
2041 b -
' D J .832 .018

 

b. Genomic structure

Mspl Hhal Dral

l l l

C/G G/A C/T

0. Observed Haplotypes

Frequency Haplotype
C G C 69.1% 111
G G T 18.2% 212
G A C 10.7% 221
G G C 1.4% 212
C G T 0.6% 112

Figure 3. SNP location, linkage disequilibrium and haplotype distribution.

Panel a shows the distance and pairwise linkage disequilibrium between the SNP markers. Panel
b Shows the genomic structure of the ADRAZA gene. The transcribed mRNA is shown as a thick
line and the portion that codes for protein is shown as a rectange. Panel c provides the nucleotide
Composition at each SNP for the observed haplotypes and frequency of each haplotype.

HaPIOtype frequencies were determined using the EM algorithm.

34

Din"

 

 

 

3550
Ptitlt
8X1).

but 1‘

Transmission disequilibrium testing using inditddual SNPS

The transmission disequilibrium test (TDT) was conducted to assess association
and linkage with each of the 3 SNPS using the parent—offspring trio data described above.
As shown in Table 2, the TDT revealed a significant association of the Dral RFLP with
the ADHD Combined subtype (p=.033). If both combined and primarily inattentive
subtypes are considered together, a p-value of .028 was found. The less common allele of
the Dral polymorphism was preferentially transmitted to ADHD children. There was no
preferential transmission of an allele of Hhal but transmission of the G allele of the Mspl
RF LP approached significance in the ADHD-(C+PI) group

The composition of our total sample, which contains non-disordered control
Children, some with intermediate symptom counts, allowed us to use tests such as the
QTDT to assess association of each of the ADRA2A SNPs with the quantitative ADHD
symptom dimensions (in addition to the diagnostic categories). We also had similar
ADHD symptom data on parents. We included all symptom data (i.e., from case and
control children as well as parents) in the QTDT analyses in order to make the symptom
distribution resemble the population distribution as closely as possible. As shown in
Table 3, both the inattentive and hyperactive-impulsive symptom scores showed
association with the Mspl RF LP and even stronger association with the Dral
POIymorphism. In contrast, neither symptom dimension was associated with the Hhal
SNP. When results were repeated excluding parental data, these associations were similar

but fell shy of significance.

35

Haplotype analysis

The finding of stronger association of ADHD subtypes and symptom dimensions
with the Dral polymorphism than with the Mspl RFLP suggests that the former may
either be closer to a functional polymorphism or may “tag” a haplotype that contains the
functional polymorphism. Therefore, the relations are determined among the alleles of

each of the three SNPs tested in this study.

Table 6. Transmission disequilibrium test results for each haplotype and ADHD subtype.

 

 

 

 

Haplotypes* ADHD type T NT RR X2 P value

ADHD-C 6 12 .50 2.00 .16

1 1 1 ADHD-P1 3 7 .43 1.60 .21
ADHD-(C+P1) 9 19 .47 3.57 .059

ADHD-C 3 5 .60 .50 .48

221 ADHD-P1 3 2 1.50 .20 .65
ADHD-(C+Pl) 6 7 .86 .08 .78
ADHD-C 11 2 5.50 6.23 .013

212 ADHD-P1 6 3 2.00 1.00 .32
ADHD-(C+P1) 17 5 3.40 6.55 .011

 

ADHD-C: Combined subtype. ADHD-PI: Primarily Inattentive subtype. ADHD-(C+PI): both of
the subtypes. T: transmitted. NT: non-transmitted. RR: relative risk. *For haplotypes: At each
position 1: common allele, 2 = less common allele. For example: 111; common allele at Mspl,
Hhal, and Dral restriction sites, 212; rare allele at Mspl, common allele at Hhal, and rare allele at

Dral restriction sites. Empirical p values are presented.

In order to capitalize on the LD among the three SNPS in ADRA2A, we next
Conducted TDT analyses using multi-marker haplotypes to determine whether this
yielded stronger results than tests perfomied with each SNP alone. These results are

Summarized in Table 6. The haplotype containing the rarer alleles of the Dral and Mspl
36

RFLPs, and the common allele of the Hhal RF LP (i.e., haplotype 212 in Figure 3c) was

preferentially transmitted to children with both subtypes of ADHD. The preferential

transmission of haplotype 212 to affected offspring likely drives the marginal

significance of haplotype 111, which is less frequently transmitted to affected children.

Similar results (summarized in Table 7) were obtained using QTDT analyses of linkage

and association between the ADHD symptom dimensions and the ADRA2A haplotypes

for the entire sample, including both case and control children and parents.

Table 7. Haplotype analysis for ADRA 2A using QTDT and ADHD symptom dimensions.

 

 

 

ADHD type Haplotypes X2(df) p-value Direction

111 600(1) .016 Decreased risk
211 NT - -

Inattention 221 0.03(1) NS -
2 I 2 10.59( 1) .001 Increased risk
1 12 NT - -
1 1 1 4.76(1) .030 Decreased risk
211 NT - -

Hyperactivity-impulsivity 221 0.06(1) NS
212 10.55(1) .004
112 NT -

‘

Increased risk

At each position 1: common allele, 2: less common allele. For example: I 1 1; common allele at

Mspl, Hhal, and Dral restriction sites, 212; rare allele at Mspl, common allele at Hhal, and rare

allele at Dral restriction sites. NT: Not Tested because of small number. NS: Not significant.

Empirical p values are presented.

37

Discussion of initial haplotype analysis

The ADRA2A gene may be an important risk factor for ADHD in light of the
role of its gene product in attention and the executive functions subserved by the
prefrontal cortex and associated circuits thought to be involved in the disorder.90 Despite
its potential relevance, only a handful of studies have investigated the ADRA2A gene as a

potential risk factor for the development of ADHD"7'48‘50'I40

. These studies analyzed an
Mspl polymorphism in the promoter of the gene and looked for association with ADHD
and/or its symptoms using a variety of statistical approaches. Comings, et al. found that
in Tourette’s Syndrome patients who also met DSM-IV criteria for ADHD, there was a
modest correlation between symptom scores and the Mspl polymorphism, but the degree
to which that sample represented the complete spectrum of ADHD patients is unclear.
There are only two studies of ADHD children without Tourette’s Syndrome, and these
yielded incommensurate results. All of these previous studies relied on a single biallelic
SNP to test for association between ADHD and ADRA2A, and thus did not adequately
sample the array of alleles in this gene. This may lead to Type II errors in assessing the
relevance of the gene to the etiology of ADHD. The present report excluded patients
with Tourette’s Syndrome and utilized a strategy of testing multiple SNPS and examining
haplotypes. Based on a survey of the literature, it is the first study to do so with this gene
in relation to ADHD. The positive results reported here therefore provide important new
evidence that the ADRA2A gene is involved in the etiology of ADHD, and further clarify
that the SNP assessed in prior studies may not be the most important marker in the gene
with respect to risk for ADHD.

Three polymorphic SNPS spanning a 3 kb genomic region were chosen for the

study of the ADRA2A gene. These SNPS are in moderate LD and define one common
38

haplotype (frequency = .69), in addition to two moderately frequent haplotypes
(frequency =. 18 and .11), in our control population (Figure 3c).

Analyses of the data using the TDT produced significant findings of association
and linkage for two of the three SNPS tested. Previous studies implicated the m allele (the
rarer G allele) of the Mspl marker in the risk for ADHD48'50‘HO. A trend was found for
association and linkage between the m allele of Mspl and ADHD using the TDT (p=.l3
for ADHD-C and p=.086 for ADHD-C+PI). In contrast, TDT analysis of the Dral RF LP
yielded significant results for ADHD-C (p=.03) but not for ADHD-PI (p=.37), as well as
for both subtypes combined (p=.028 for ADHD-C+PI).

Despite the identification of ADHD as a categorical disorder in DSM-IV,
quantitative genetic analyses suggest that ADHD symptom dimensions show similar
genetic influences at all levels of severity.‘33 These findings warrant consideration of the
association of candidate gene markers with dimensional symptom ratings. We conducted

141.142

such analyses using the QTDT. This approach revealed significant association and
linkage of symptoms of inattention with the rare allele of the Mspl polymorphism
(p=.022) and confirmed association and linkage with the rare allele of Dral (p=.003).
Similar findings were also obtained with these alleles and the hyperactive-impulsive
symptom dimension (p=.037 for Mspl and p=.015 for Dral). The results of the QTDT
analyses suggest that the functional risk-inducing allele might reside on chromosomes
containing the rare alleles for both Dral and Mspl.

The TDT was repeated in order to evaluate association and linkage between
ADHD subtypes and symptom dimensions and specific ADRA2A gene haplotypes. The

haplotype containing the rare alleles of both the Dral and Mspl markers was significantly

associated with ADHD and the combined subtype. The QTDT results suggested that the
39

same haplotype was associated with severity on both the inattentive and hyperactive-
impulsive symptom dimensions. These results suggest that the rare allele of Dral may be

closely linked to a functional polymorphism in the ADRA2A gene.

Conclusion

This haplotype study emphasizes several points concerning the search for
functional alleles of genes that contribute to the inheritance of complex disorders and
diseases. It is clear that marker selection in the candidate gene should not be limited to a
single polymorphism. If the Hhal polymorphism had been the only marker selected for
analysis, no indication of a significant association between ADHD and the ADRA2A gene
would have resulted. In general, it is only when a functional polymorphism is being
tested that a single marker will yield the most significant results. When the functional
polymorphisms are not known, as is almost always the case, it is prudent to identify
several polymorphisms in the candidate gene and to test each for association, both singly
and in combination using the haplotypes that they constitute. In the case of ADRA2A, we
identified three common haplotypes in the gene, the analyses of which allowed us to
better demonstrate association and linkage between the ADHD subtypes and symptom
dimensions and ADRA 2A.

It is important to be aware that the power of the TDT depends not only on effect
size and the mode of inheritance, but also on the allele frequency of the SNP in the
population. In the present study, given our sample size we had adequate power to detect
the largest effects in the ranges seen in our data within a multiplicative model for Dral
and Mspl, but we had much lower power to detect these effects for Hhal, given both its

smaller effects and its much greater allele frequency. Furthermore, power was

40

considerably greater for the haplotype analyses than for the analyses of each individual
SNP, underscoring the value of incorporating multiple markers in studies of association
with candidate genes.

There are several limitations of this study that should be noted. Aside from the
relatively small sample, the most important of these is the reliance on parental symptoms
in the QTDT analyses. These retrospective symptom ratings are vulnerable to multiple
biases, even when obtained in a careful structured interview as in our study. Therefore the
quantitative results should be viewed with some caution until replicated. Nonetheless, we
note that in our sample population the QTDT analyses relied only on allelic transmissions
from parents and used the parents’ symptom scores solely for the purpose of estimating
the population mean of the ADHD symptom dimensions.

In conclusion for initial haplotype analysis, our results suggest that ADRA2A is
associated and linked with ADHD and that the functional polymorphism is closely linked
to Dral. Although ambiguous haplotypes were not included, the haplotype TDT and
QTDT results suggest that the functional allele is likely to be a less frequent allele (~.20)
and is present on the “212” haplotype, which represents the rare alleles of Mspl and Dral
and the common allele at Hhal. It is possible that there is more than one functional
polymorphism within this haplotype that contributes to the gene’s effects. These results

underscore the potential importance of noradrenergic systems in the etiology of ADHD.

Materials and methods
DNA Preparation
For buccal DNA preparation, a modified method described by Meulenbelt was

performed in which cheek swabs were used as samples followed by DNA preparation

41

using phenol/chloroform purification (average 60 ug DNA per collection) using the
following procedurem.
l. Swabs collected in 5 ml of an STE solution were centrifuged at 2000 rpm in 50 ml
tubes for 5 minutes.

2. Swabs, briefly suspended by a pulse vortex are inverted and transferred into a new 50
ml tube, and centrifuged at 2000 rpm for 5 minutes.

3. Swabs were pulled out and discarded using a clean glove after each set.

4. The liquid briefly suspended by a pulse vortex is transferred to a labeled 15 ml tube
containing 5 m1 of phenol/chloroform (1:1).

5. The tubes were inverted gently 10 times, and allowed to settle for 5-10 minutes.

6. The tubes were centrifuged at 4000 rpm for 8-10 minutes.

7. The aqueous layer of the solution was transferred to a new 15 ml tube containing 5 ml
of chloroform, and the tubes were inverted gently 10 times.

8. The tubes were centrifuged at 4000 rpm for 8-10 minutes.

9. The aqueous layer of the solution was transferred to a new 15 ml tube containing 5 ml
of 2-propanol, and the tubes were inverted 50 times.

10. The tubes were centrifuged at 4000 rpm for 8-10 minutes.

11. To leave the pellet, the solution was poured off carefully into a beaker, and 2 ml of
70% ethanol was added to the pellet. (After this procedure, the tubes could be stored in a
refrigerator for overnight.)

12. The tubes were centrifuged at 4000 rpm for 5 minutes.

13. Ethanol was poured off carefully so that the pellet did not slide out from the tube, and
the tubes were inverted onto clean papers to allow the pellet 20-30 minutes of drying time.

14. 200 111 of DNA hydration solution was added to the pellets, and let the tubes were left
42

to sit overnight at room temperature.
15. To dissolve the pellet, the solution was suspended by pulse vortex, and transferred to
screw—capped tubes for storage at —20 °C.

The concentrated stock was diluted (1/ 100 or 1/50), and the diluted solution is
read by UV spectrophotometer at 260 and 280 nm for the measurement of rough DNA
concentration and purity.

Genotyping

Eight polymorphisms were studied from the three candidate genes by PCR and
restriction fragmentation.

For DRD4, two well-replicated polymorphisms in ADHD, insertion/deletion
promoter polymorphism and variable number of tandem repeats (VNTR) in exon 3, were
selected and assayed with minor modificationsm. The DRD4 120-bp tandem repeat
polymorphism was assayed in 20 111 reaction mixture containing 20 ng of genomic DNA,
200 uM dNTPs, 1 1.1M of each primer, 1.5 mM MgC13, IX PCR buffer, and 0.5 units of
'Taq DNA polymerase with the same primer sets (5’-GTTGTCTGTCTTTTCTCA
TTGTTTCCATTG-3’ and 5’-GAAGGAGCAGGCACCGTGAGC-3’). Amplification was
conducted under the following conditions with a hot start; an initial denaturing step at
94 °C for 3 minutes followed by 35 cycles consisting of 30 seconds at 94 °C, 30 seconds
at 61 °C, and 1 minute at 72 OC, and the final extension step for 5 minutes at 72 0C using
A81 9700. The VNTR was amplified in 25 ul reaction mixture containing 100 ng of
genomic DNA, 200 11M dNTPs, 0.5 uM of each primer, 1X Q solution, 1X Q PCR buffer,
and 0.625 units of Taq DNA polymerase using primer sets, 5’-
CGTACTGTGCGGCCTCAACGA-3’ and 5’-GACACAGCGCCT GCGTGATGT-3’.

The DNA was amplified with a hot start procedure including an initial denaturing step of

43

30 seconds at 96 °C, followed by 40 cycles consisting of 30 seconds at 95 °C and 90
seconds at 68 °C, and the final extension step for 4 minutes at 72 0C. After the
amplification, the DNA was detected in 1.5 % argarose gel for the 120-bp tandem repeat
polymorphism and 1.2% argarose gel for VNTR stained with ethidium bromide.

Three polymorphisms for SLC6A3, exon 9 SNP, intron 9 SNP, and VNTR in
exon 15, were selected and typed with minor modification as described previously”.
VNTR was amplified using the primer sets, 5’-ACTCCTTGAAACCAGCTCAG-3’ and
5’-TATTGATGTGGCACGCACCT-3’ in the reaction mixture containing 20 ng of
genomic DNA, 62.5 11M each dATP, dTTP, dCTP, 31.25 11M dGTP, 31.25 11M deaza
dGTP, 1 11M of each primer, 1.5 mM MgC12, 1X PCR buffer, and 0.5 units of Taq DNA
polymerase using deaza dGTP as describedmg. The procedure includes an initial
denaturation for 3 minutes at 95 °C, is followed by 35 cycles consisting of 30 seconds at
95 °C, 30 seconds at 58 °C, and 45 seconds at 72 OC, and the final extension step for 2
minutes at 72 °C. For amplification of the other two SNPS, the PCR used 60 ng of
genomic DNA, 200 11M dNTPs, 1 uM of each primer, 1.5 mM MgC12, 1X PCR buffer,
and 0.5 units of Taq DNA polymerase (primer sets: 5’-CACAGCGTGGGCTCTGTG-3’
and 5’-GGTGGAAGGAACCCAACTG-3’ for the exon 9 SNP and 5’-
GTCGTGCCGCCAT AGAAG-3’ and 5’-CTGCACACAGAGGACAGGGT-3’ which is
mutated from the original sequence in the genome for a proper restriction cut for the
intron 9 SNP). The cycling parameters involve an initial denaturation for 4 minutes at
94 °C; 35 cycles consisting of 40 seconds at 94 °C, 40 seconds at 65 0C for exon 9 SNP
and 57 0C for intron 9 SNP, and 30 seconds at 72 °C; and the final extension step of 5
minutes at 72 OC. The amplified DNAs are digested by 10 units of restriction enzymes at

37 0C overnight (Ddel for exon 9 and PflFI for intron 9). For the efficiency reason, lul of
44

buffer 4 was added to the restriction digestion of the amplicon of intron 9. The DNA was
detected in 1.5 % argarose gel for the VNTR polymorphism and 3% argarose gel for two
other SNPS stained with ethidium bromide.

Three SNPS in ADRA2A were selected based on their spacing and frequencies as

described previously143 . The promoter SNP, rsl80044 (Mspl RFLP), was typed by a

109

modified amplification of the region using deaza dGTP as described . Briefly,

polymerase chain reaction (PCR) was performed in 20 pl reaction mixture containing 40
ng of genomic DNA, 62.5 uM each dATP, dTTP, dCTP, 31.25 uM dGTP, 31.25 uM
deaza dGTP, 1 uM of each primer, 1.5 mM MgC12, IX PCR buffer, and 0.5 units of Taq

DNA polymerase. For the 5’ UTR SNP, r5180045 (Hhal RFLP), PCR amplification were

performed using primer sets, 5’-CCAAGTTATCAGGCCACCGA-3’ and 5’-

TGCTCCTGGCGGAACAT GAA-3’ in 20 111 volume containing 40 ng of genomic DNA,
200 11M dNTPs, 1 11M of each primer, 1.5 mM MgC12, 1X PCR buffer, 2 1.11 DMSO, and
0.5 units of Taq DNA polymerase. Amplification included an initial denaturing step at
94 °C for 3 minutes followed by 35 cycles consisting of 30 seconds at 94 0C, 30 seconds
at 60 °C, and 45 seconds at 72 °C, and the final extension step of 5 minutes at 72 °C.
After amplification, 10 units of Hhal restriction enzyme were added and digestion was

performed at 37 °C for 2 hours. The region for the 3’UTR SNP, r8583668 (Dral RFLP),

was amplified in 20 ul volumes containing 40 ng of genomic DNA, 200 11M dNTPs, 1

pM of each primer, 1.5 mM MgC12, 1X PCR buffer, and 0.5 units of Taq DNA

polymerase (primer sets: 5 ’-TACAAGGGCATGGCTCACAA-3’ and 5 ’-

CCAAGGCCAGGATTTCAACA-3’) using the same cycling parameters as above.

Digestion of the PCR product was performed with 10 units of Dral restriction enzyme at

45

37 °C for 2 hours. All restriction fragments were detected using 3% agarose gel stained

with ethidium bromide.

Data Analysis

Hardy-Weinberg Equilibrium tests were performed using contingency tables.
Case-control association was tested for each SNP. An increased alpha-level was
considered using primarily the level of p= .01 to establish statistical significance in view
of the number of statistical tests conducted to reduce the familywise Type I error rate,
while preserving sufficient power to avoid excess Type 11 error. All statistical tests were
two-tailed.

For within-family analyses of association and linkage between each of the

ADRA2A SNPS and the ADHD diagnostic subtypes, we used the original TDT (i.e., a
McNemar’s chi-square test of biased transmission of alleles from heterozygous parents to
their affected offspring)'44. The quantitative TDT was performed using QTDT
software'4"'42. Because of the very different distributions in ADHD symptom dimension
scores between parents and their offspring, the polygenic variance (05) as well as the
additive genetic variance (of) in the QTDT could not be calculated. This resulted in p-
values that were very similar to empirical p-values calculated from 1,000 permutations.

One non-mendelian family (probably due to sample mix of the family) was not

included in data analysis. In case of a strong suspicion of non-partemity of second child,
the second child’s genotype data was eliminated from the analysis. However, if the non-
parternity corresponds to the first child who is phenotyped, the father’s genotype data
was deleted from the analysis.

Parents of controls were used to determine population haplotype frequencies.

Both the use of manual procedures and the expectation maximization (EM) algorithm via

46

maximum likelihood estimation produced the same results for haplotype estimation. For
analyses using haplotypes, where phase was ambiguous, the trios were omitted from the
analysis (i.e., in 13 of 177 family samples). Linkage disequilibrium among the SNPs in
ADRA2A was estimated using the GOLD software package“. We report findings
separately for ADHD-C and ADHD-PI, as well as pooled results for both subtypes, in

view of disagreement in the field about the degree of their etiological similarity and

whether or not their results should be pooledI45 .

47

Chapter 3

Phenotypic Consideration

Introduction
Unlike single gene disorders, complex phenotype presents much more difficulties
in studying complex traits. Most complex traits include some degree of comorbidity and
subtypes of the disorders. Although each disorder has a main pathophysiology, it seems to
overlap at least partially with other similar disorders and the resulting symptoms can be
somewhat distinguished depending on the clinical presentation. Complex traits can be
frequently found in mental disorders due to the complex network of the brain.

As expected, ADHD often occurs in conjunction with other major psychiatric
disorders. Common comorbidity disorders and their relative frequencies are as follows:
oppositional defiant disorder (33%), conduct disorder (25%), anxiety disorders (25%),
depressive disorders (20%), and learning disabilities (22%)8. The comorbidity of those

disorders suggests an overlapping pathophysiology and possible genetic etiology with
ADHD. It is not clear that the overlapping pathophysiology does mean stronger genetic
influence of the comorbid genetic locus. It is worthwhile to examine the comorbidity and
distinguish the etiology, but, in the current stage of the genetic association studies, to find
out each genetic etiology of the disorder seems more appropriate.
In this chapter, to find out if a polymorphism is associated with ADHD
phenotype, DSM-IV based case-control and TDT test are primarily considered including
their subtypes. A total of 228 nuclear families that is slightly more than were used in
chapter two were studied in three groups: Non-ADHD control (n=70), ADHD-combined

ype (n=95), and ADHD-primarily inattentive type (n=29) with 64.4% Caucasians and
48

67% boys. This sample population is a balanced collection between cases of controls of
several ethnic groups (Caucasian, African American, Hispanic, Asian, American Indian,
and mixed others) with p-value of .57, so that the case-control association is not affected

by population stratification. The final sample included children aged 6-13 years (mean =

9.6). To address concerns in regard to ADHD phenotypes, several relevant

endophenotypes were tested also. Finally, one important yet unexplained feature of

ADHD, the gender difference, was addressed.

Similar to those of chapter two, the children were recruited via a community-
based, multi-gate strategy in which more stringent diagnostic procedures were applied at
each stage in order to establish cases. In the first stage, common rule outs were identified
such as autistic disorder, mental retardation, neurological disease, and sensorimotor
handicap. In the second stage, parent and teacher normative ratings were obtained to
make sure the child had elevated levels of behavior problems in both settings (for
potential ADHD participants) or had normal range behavior across settings (for potential
control participants). During the final stage, a structured diagnostic interview (the NIMH
Diagnostic Interview Schedule for DSM-IV) was performed with the primary caregiver
to establish that full DSM-IV criteria were met for the ADHD groups and that Control
children did not have ADHD. Parent and teacher data were combined in an “or”

algorithm to arrive at the final symptom count in assigning the ADHD subtype. Thus, if a
symptom was endorsed by the parent on the DISC-1V, or was rated by the teacher as a
“2” or a “3” on the 0-3 scale used to rate the items ADHD Rating scale, it was counted as
present. At this final stage, we also assessed other psychopathologies, ruling out children

with Tourette Disorder, bipolar disorder, psychosis, or learning disability and recording

other comorbid conditions for secondary data analysis.
49

DSM IV based associations.

As described previously, there are three subtypes of ADHD. Two relatively

common subtypes, ADHD-C and ADHD-PI types, appear to differ both cognitively and

with regard to familial history65‘66‘l46. Although neuropsychological data has not

differentiated them'47‘l48, there is a suggestion that the two subtypes are entirely different

disorders”. Moreover, there is evidence that the comorbid symptoms with other
disorders differ depending on the subtypesm.

The concern, that the ADHD-combined type (ADHD-C) and ADHD-
predominantly inattentive type (ADHD-PI) are totally different subtypes of ADHD, leads
to consider segregated analysis between two subtypes. Although previously discussed
that ADHD combined and inattentive subtypes may be distinct conditionsm, it is also
noted that there is an argument that these two ADHD subtypes commonly share many
neuropsychological features'so. Also, the inattentive type may to a large extent represent a
milder version of the ADHD combined typel5 1. Therefore, the tests were conducted on
both ADHD-C and ADHD-PI groups, as well as the combination of the two for purposes
of the present paper to maximize power. TDTs and case-control tests were conducted on
those groups.

In Table 8, ADHD-PI shows generally reduced association probably due to the
smaller sample size. There is no major difference in the trend of associations between
ADHD-C and ADHD-Pl. There are some differences in association levels between these
two groups, but it is hard to surmise further due to the small sample size. With the similar
trend between ADHD-C and ADHD-PI, the combined grouping of both and ADHD-C

also show similar results. More sampling on ADHD-Pl seems necessary for further

50

speculation. These results indicate that, at least in our sample population, the genetic

etiology of ADHD in three genes may be similar.

Table 8. Genotype Association Results (p-values of chi-square test).

 

 

 

 

Control vs Control vs Control vs
Gene Polymorphism
ADHD-C + ADHD-Pl ADHD-C ADHD-PI
Insertion/deletion .003 .003 .20
VNTR .76 .99 .18
Exon 9 .82 .67 .88
SLC6A3M 1618119” _ V. 046 . , . .06 . . “2.111. ..
rs 1 800544 (Mspl) .52 .40 .98
ADRA2A rsl800545(HhaI) .. 43 ., .. “-55.. V. .,. _, V, V 52.. V.
r5553668(DraI)099 . .. .4079 58

 

Exact numbers for ADHD-C + ADHD-PI are indicated in Tables 13-15.

Although our sample population is balanced with respect to cases and controls in
ethnic groups, the same case-control tests using only Caucasians were tested to address if
there is ethnic-specific association. Overall, the results (Table 9) were not much different
from the previous results in Table 8. Some changes in Caucasian only associations are
summarized below; the association of insertion/deletion polymorphism in DRD4 was

reduced and the association of VNTR in SLC6A3 was enhanced. One interesting feature
is that the association of VNTR was strongest in the case-control test using only ADHD-
PI although it is smaller sample size. It is noteworthy that, as described in the previous
chapter, this association is in the opposite direction compared to the research of other

groups. Further discussion regarding this question is addressed in Chapter 5. The

51

Caucasian-only association suggests that there may be some ethnic differences in the
level of association of each polymorphism depending on the LD with functional

polymorphisms, although further research using increased samples is necessary.

Table 9. Genotype Association Results (p—values of chi-square test on only Caucasians).

 

 

 

 

. Control vs Control vs Control vs
Gene Polymorphism
ADHD-C + ADHD-PI ADHD-C ADHD-PI
Insertion/deletion .062 .13 .16
VNTR .31 .28 .56
Exon 9 .85 .90 .71
SLC6A3 Intron9 054 .. 073 .ll ,
rs 1 800544 (Mspl) .84 .71 .79
ADRAZAVLM-rsl800545(Hhal)30 .. 1.1. . .. . 7.2
rs553668(DraI) 21. , ., H.039 55

 

TDT shows a pattern similar to that described in the previous chapter with
smaller samples (Table 10). None of polymorphisms in DRD4 and SLC6A3 is
significantly associated with ADHD using TDT, and again the Dral RFLP in ADRA2A
shows borderline significance reduced a bit more than the result in the previous chapter.
As noted in Table 8, slightly increased sample size produced more significant association
of Dral RFLP in ADRA2A using the case-control test. It appears that the detection of
association is moved from TDT to case-control test in ADRA2A after adding a few more
samples. This result is probably due to the fact that both the added ADHD subjects and
their parents do have more risk alleles of Dral. Overall difference patterns between

subtypes are similar to the case-control association studies.

52

Table 10. Transmission disequilibrium test results for each ADRA2A SNP and ADHD

 

 

 

 

subtype.
SNPs ADHD type T NT RR X2 P value

ADHD-C 17 9 1.89 2.46 .12

Mspl (G allele) ADHD-P1 10 6 1.67 1.00 .32
ADHD—(C+PI) 27 15 1.80 3.43 .06

ADHD-C 5 7 .71 .33 .56

Hhal (G allele) ADHD-P1 4 2 2.00 .67 .41
ADHD-(C+Pl) 9 9 1.00 0 1.00
ADHD-C 14 5 2.80 4.57 .039

Dral (T allele) ADHD-P1 8 6 1.33 .82 .59
ADHD-(C+PI) 22 11 2.00 4.84 .056

 

ADHD-C: combined type. ADHD-P1: primarily inattentive. ADHD-(C+P1): both ofthe types. T:

transmitted. NT: non-transmitted. RR: relative risk.

Recognizing that ADHD is a problem of degree rather than existence and that
this sample population contains many controls and some intermediates, the quantitative
TDT measurement of ADHD is employed. There are nine symptoms of inattention or
hyperactivity-impulsivity on DSM-IV criteria that defines ADHD if an individual has at
least six symptoms. The number of symptoms is used for quantitative analysis for the
association. Recognizing that 02g and (528 are not calculated using parents’ and their
children’s symptoms in our sample population which is the same as the smaller samples
of the previous chapter, the variance component tests were not pennutated.

Same as TDT, the QTDT tests did not demonstrate any association with
polymorphisms in DRD4 or SLC6A3. However, as summarized in Table 11, ADRA2A

shows an association with ADHD. Again, these associations are reduced compared to the

53

smaller samples in the previous chapter. Given the importance of quantitative
measurements, the QTDT using symptom counting is robust in replacing TDT especially
in our sample population, which has controls as well as subjects with intermediate

symptoms.

Table 11. QTDT results for ADHD symptom dimensions associated with ADRA2A

 

 

 

 

 

 

 

polymorphisms.

ADHD type Marker Allele* X2(df) P
Mspl G 281(1) .09

Inattention Hhal A 0.23(l) NS
Dral T 4.76(1) .029
Mspl G 2.24(]) NS

Hyperactivity-impulsivity Hhal A 0.62(1) NS
Dral T 405(1) .044

 

NS: not significant.

*The allele conferring increased risk is denoted.

Endophenotypes

Attention deficit hyperactivity disorder (ADHD) is a common, costly'sz, and
impairing'50 condition typically diagnosed in childhood. It evidences substantial
heritability153 and as a result, molecular genetic studies have proceeded rapidly, with
replicated findings for several candidate genes's". Recognizing the measurement of clear
phenotype, the endophenotype, quantifiable and intermediate constructs, becomes the
immediate interest in the genetic research of ADHD. Therefore, finding the measurable
phenotypes is an important step in the genetic study“. The appropriate endophenotypes

can be: I) correlated with ADHD symptoms or disorder; 2) amendable to objective,

54

ideally quantitative, measurement; 3) present in relatives; and 4) theoretically or
empirically related to the etiology of the disorder. The expectation is that endophenotypes
involve the same biological pathways as the disease but are nearer the relevant gene
action. In this study, the neurocognitive measures were focused on due to their theorized
mediating causal role in ADHD.

Among the relevant neurocongitive circuits of ADHD, evidence from
neuroimaging studies converges on the involvement of frontal-striatal-thalamic
dysfunction in ADHD with involvement of the prefronatal cortex, basal ganglia,

33"56. To date, the primary theoretical and empirical

cerebellum and corpus collosum
emphasis has relied on catecolamine transmission, partly because this is presumed to be
the site of psychostimulant action used in treating ADHD.

Depending on the neuropsychological basis, possible neurocognitive
endophenotypes can be found in executive functions and inhibitory controls. One such
possible endophenotype of ADHD is a difficulty to suppress non-goal motor response or
competing response tendencies. The frontal-subcortical—thalamic circuits are important in
suppressing competing behavior as well as in representing intended behavior. Operational
definitions of behavioral suppression have led to several candidate endophenotype
measures, of which perhaps the most well established is the Logan stop taskm.

Motor inhibition or response suppression is operationalized with the tracking
version of the stop task, a computerized choice-reaction time task using the same
procedures as Logan et al.157 and Nigg15 8. In the tracking version of this task, stop signal
reaction time, the index of inhibitory control, is estimated by subtracting mean stop signal

. 157 . . . . .
latency from mean go response time . Go response time and variability of response time

serve as indices of regulation of arousal, activation, or effort. Each of the outcome

55

variables is mean reaction times with excellent reliability.

In this study, the tracking version of the stop task was used, which is supported
by recent data as providing the most robust assessment of stop signal reaction time (stop
signal RT)'59. Generally, ADHD children have deficits on the speed and variability
measures. Both stop signal reaction time and response variability show the strongest
correlations between probands and relatives among several candidate measurements.
Recognizing the nature of those endophenotypes, the quantitative analysis was primarily

tested using the QTDT software.

Table 12. Endophenotype analyses (p-values of tests for total evidence of association and

 

 

 

 

 

QTDT).
. Stop Signal RT Response variability
Gene Polymorphism
Association TDT Association TDT
Insertion/deletion NS NS NS NS
VNTR NS NS NS NS
Exon 9 NS NS NS NS
SLC6A3 .. Intron9 .. 0569NS .. . NS .. NS V
rsl800544 (Mspl) NS NS NS NS
ADRA2A r31800545(HhaI)0157NS NS _ . . NS ..
r5553668(Dra1) . NS NS UNS- “NS .........

 

NS; not significant with p-values higher than .10.

Tests of quantitative associations and transmission disequilibrium were
summarized in Table 12. Because parent phenotypes are not included, all the variances

are the same in those analyses. There is no significant population stratification except the

56

polymorphism on Exon 9 of SLC6A3 on the response variability (x2 = 4.84, p-value
= .0278). Unlike association results using DSM-IV based phenotypes, most of the tests
fail to find associations between those endophenotypes and polymorphisms. One
interesting result is the association between Hhal RFLP on ADRA2A and stop signal
reaction time. Recognizing that the Dral RFLP on ADRA2A is associated with DSM-IV
based phenotypes, this endophenotype may be associated with the gene, ADRA2A, in
different ways than the DSM-IV phenotypes.

Although candidate endophenotypes are familial and correlated with ADHD, a
cognitive endophenotype may be partially overlapped with ADHD so that the children
without ADHD also have the endophenotype. It is possible that a genetic study on the
endophenotype can dilute the pure genetic effect on ADHD. Another concern on the
endophenotypes is the endophenotype may contain other sets of genes affecting the
endophenotype. As an example, the stop signal task may include the genes related to the
peripheral nervous system, although ADHD mostly concerns the central nervous system.
This result suggests that the endophenotypes may need to be cautiously selected for the
genetic association study only if they are proven to be heritable and to be exact subsets of
ADHD or biologically relevant to ADHD depending on neuroimaging or

neurophysiology.

Gender difference
Under-examined in the genetic literature, like most other psychopathology,
ADHD does not afflict boys and girls equally. In childhood, the ratio of affected boys to

girls ranges from 3:1 to 8:1, depending on whether one surveys community or clinical

samplesmo. It has been unclear whether this difference in prevalence reflects merely

57

61

differences in socialization or diagnostic detection of boys versus girlsl or whether girls

75.162

are protected from the risk factors that cause the disorder in boys . Supporting this
latter perspective have been meta-analytic findings that although girls with ADHD tend to
have less severe behavioral disturbance than boys in terms of less severe hyperactivity
and conduct problems, they may have greater intellectual impairment7‘74. One
neuroimaging study found more marked hypoactivation of brain region(s) in girls than
boys with ADHD'“, although results need replication. Nigg et al. found that girls with
the ADHD inattentive type had deficits in response inhibition that were not observed in
boys with the inattentive type of ADHD“. Numerous neurodevelopmental and
behavioral differences between boys and girls, rooted in their obvious differential

165.166
, can be noted to

hormonal exposures during prenatal development and subsequently
support the possibility of distinct neurobiological mechanisms underlying ADHD in boys
versus girls.

This possibility has already been noted in other areas of psychopathology. For
example, protective effects of estrogen may account for delayed onset and reduced
severity of schizophrenia, whereas estrogen withdrawal may be implicated in an
increased incidence of Alzheimer’s disease in post menopausal womenm'ms. However,
unlike schizophrenia or Alzheimer’s, ADHD begins in early childhood, before significant
sex differences in circulating estrogen arise. Although the disorder can persist into
adulthood, the gender ratio in adults is unclearmg. Thus, hormonal effects at puberty
could still affect ADHD prevalence later in development. Even so, childhood differences
in expression of ADHD may be attributable to prenatal hormonal effects on neural
170

organization

With regard to candidate neural systems and genes, ADHD is widely suspected

58

of involving catecholaminergic dysfunctionzs‘zc’. Both dopaminergic and noradrenergic
systems are thought to be important. Gender differences in regulation of these systems
would not be unprecedented; for example, gender differences have been noted in relation

84.171.172

to cardiovascular disease both in the adrenergic system and in animal studies of

76. 7 - -
7 . One noradrenergic candidate gene

genetic correlates looking at quantitative trait loci
for ADHD, the a-ZA-adrenergic receptor (ADRA2A), apparently shows gender specific
differences in response in relation to vasoconstriction levels in animalsy’.

Also suggestive is that gender differences have been noted in physiological
response to cocaine and methamphetamine, which are dopaminergic agonists closely

related to methylphenidate, a common treatment for ADHDm“74

. Striatal dopamine
concentrations of methamphetamine treated female mice were significantly less depleted
than those of identically treated male mice. Evidence suggests that women may have a
higher synaptic concentration of dopamine in the striatum than men”. Furthermore,
fluctuating ovarian hormones cause periodic variation in the expression of dopamine

- , 7
receptors in femalesg” 6

. Moreover, the mRNA level and density of the dopamine
transporter are significantly higher in female rats than in males”. In human studies,
SPECT indicates a significantly higher density of the dopamine transporter in the striatal
region of females than of malesm, and such effects have been hypothesized as potential
mechanisms in ADHD’s gender differencesm. Animal studies suggest that such
differences may be related to female gonadal steroids regulating gene expressionm‘go.
Therefore, we sought to evaluate whether the three candidate genes selected in this study
show differential associations to risk in boys versus girls.

DRD4.

In Table 13, the association between polymorphisms in the DRD4 gene and
59

ADHD is examined. In the total sample, individuals, who are homozygous for the
insertion in the DRD4 promoter, were at an increased risk for ADHD. This was a very
significant finding in boys but not significant in girls, although girls did show a similar
trend. No significant association with the DRD4 VNTR was found in our sample
population.

When we excluded non-Caucasians from the data set and recalculated the
association between DRD4 I/D and ADHD, the trends were very similar but, probably
due to the reduced sample size, did not reach greater significance (boys: p= .02; girls:
p= .42; overall: p=.06). No significant difference between boy and girl controls or cases
was found (control: p = .15; cases: p=.46).

SLC6A3.

Three polymorphisms were similarly studied for SLC6A3 (Table 14). The
haplotypes using these polymorphisms were previously associated with ADHD”. No
significant associations were found for the exon 9 SNP. For the intron 9 SNP in SLC 6A 3,
a trend was found in the total cases vs. controls for increased risk in the individuals
carrying the G allele. When the genders were considered separately, girls had a very
significant ADHD risk associated with the G allele; boys did not. When only Caucasians
were tested, similar results were obtained (boys: p= .74; girls: [F .003; overall: [F .054).
Comparison of boy and girl controls or cases yielded no recognizable significance
(control: p= .044; cases: p=.75).

The VNTR found in the 3’UTR of SLC6A3 has frequently been associated with
ADHD36'4', but this study found no significant association between ADHD and the
VNTR in either total cases or boys alone. When we examined the association using only

girls, we found a significant ADHD association with the “not 10” allele, which was also

60

seen when we restrict our sample to only Caucasians (total: p= .046; boys: p= .82; girls:
[F .003). The difference between boy and girl controls or cases did not reach significance
at our reduced alpha level (control: p= .039; cases [F .51).
ADRA2A.

Three polymorphisms found in the ADRA2A gene were tested (Table 15). For the
Mspl RF LP, no significant association was found in the total sample or boys alone case-
control groups, although the effect was near significance for girls. Restricting our
samples to only Caucasians showed no significance (total: p= .83; boys: p= .89; girls:
p= .16). The Hhal RFLP produced no significant associations in any group. Our previous
study found ADHD linked to the Dral polymorphism in the 3’UTR of ADRA2A using
TDT analysism. In the present study, which contains additional samples, we found
similar results (Table 15 footnote), and in addition, we found that the case/control
comparison approached significance in the total sample. However, there was a strong
association of ADHD with the Dral polymorphism in girls. Similar findings were seen in
the Caucasian-only subset (total: [F .25; boys: p= .72; girls: p= .002). The rarer T allele
of Dral polymorphism appears to confer risk to develop ADHD primarily in girls.
Examining girls versus boys with ADHD yielded significance (p= .009), while boy

versus girl controls did not (p=.027) considering the significance level as p-value of .01.

61

Table 13. Case-control association of DRD4 (controls vs ADHD all types).

 

 

 

 

 

 

 

 

 

Polymorphism Sample No of individuals with genotypes p-value
(percentages of genotype)
Insertion/ I/I l/D D/D
Deletion Total Controls 33(48.5) 30(44. 1) 5(7.4) .003*
Cases 89(73.0) 26(21.3) 7(5.7)
Boys Controls 22(50.0) 20(45.5) 2(4.5) .002*
Cases 63(75.9) l4(16.9) 6(7.2)
Girls Controls ll(45.8) 10(41.7) 3( 12.5) 14
Cases 26(66.7) 12(30.8) 1(2.6)
VNTR ~7/~7 ~7/7 7/7
Total Controls 38(55. 1) 28(40.6) 3(4.3) .76
Cases 73(60.3) 44(36.4) 4(3.3)
Boys Controls 20(44.4) 23(5 1.1 ) 2(4.4) .16
Cases 52(61.9) 29(34.5) 3(3.6)
Girls Controls l8(75.0) 5(20.8) 1(4.2) .28
Cases 21(56.8) 15(40.5) 1(2.7)

 

Allele frequencies for Insertion/deletion polymorphism (l = .71, D: .29), for VNTR (7 repeat

= .25, not 7 repeat = .75). No deviation from Hardy Weinberg equilibrium was found in the total

controls.

62

Table 14. Case-control association of SLC6A3 (controls vs ADHD all types).

 

 

 

 

 

 

 

 

 

 

 

 

 

Polymorphism Sample No of individuals with genotypes P-value
(percentages of genotype)
Exon9 SNP A/A A/G G/G Sum
Rs6347 Total Controls 39(55.7) 27(38.6) 4(5.7) 70 .82
Cases 66(53.2) 48(38.7) 10(8.1) 124
Boys Controls 22(48.9) l9(42.2) 4(8.9) 45 .66
Cases 48(56.5) 29(34.1) 8(9.4) 85
Girls Controls 17(68.0) 8(32.0) 0(0.0) 25 .16
Cases 18(46.2) 19(48.7) 2(5.l) 39
Intron9 SNP A/A A/G G/G Sum
PflFI RFLP Total Controls 56(80.0) 13(18.6) 1(1.4) 70 .046
Tth1111 RFLP Cases 78(62.9) 42(33.9) 4(3.2) 124
Boys Controls 32(71.1) 12(26.7) l(2.2) 45 .74
Cases 55(64.7) 27(3l.8) 3(3.5) 85
Girls Controls 24(96.0) l(4.0) 0(0.0) 25 .005*
Cases 23(59.0) 15(38.5) 1(2.6) 39
VNTR 10/10 10/~10 ~10/~10 Sum
Total Controls 42(60.0) 23(32.9) 5(7. 1) 70 .25
Cases 58(483) 54(45.0) 8(6.7) 120
Boys Controls 22(48.9) l9(42.2) 4(8.9) 45 .94
Cases 42(51.2) 34(41.5) 6(7.3) 82
Girls Controls 20(80.0) 4(16.0) 1(4.0) 25 .010*
Cases 16(42.1) 20(52.6) 2(5.3) 38

 

Allele frequencies: Exon9 SNP (A=0.75, G=0.25), Intron9 SNP (A=0.90, G=0.10), VNTR (10
repeat = 0.76, not 10 repeat (~10) = 0.24). No deviation from Hardy Weinberg equilibrium was

found in the controls.

63

 

Table 15. Case-control association of ADRA2A (controls vs ADHD all types).

 

 

 

 

 

 

 

 

 

 

 

 

 

Polymorphism Sample No of individuals with genotypes P-value
(percentages of genotype)
Mspl RFLP C/C C/G G/G Sum
Rs 1 800544 Total Controls 33(485) 26(38.2) 9(13.2) 68 .52
Cases 50(4l.0) 57(46.7) 15(12.3) 122
Boys Controls 20(45.5) 18(40.9) 6( 13.6) 44 .95
Cases 40(47.6) 34(40.5) 10(11.9) 84
Girls Controls l3(54.2) 8(33.3) 3(12.5) 24 .072
Cases 10(26.3) 23(60.5) 5( 13.2) 38
Hhal RFLP G/G G/A A/A Sum
Rs 1 800545 Total Controls 50(71.4) 17(24.3) 3(4.3) 70 .43
Cases 96(77.4) 26(21.0) 2( 1.6) 124
Boys Controls 33(73.3) 9(20.0) 3(6.7) 45 .46
Cases 67(78.8) 16(18.8) 2(2.4) 85
Girls Controls 17(68.0) 8(32.0) 0(0.0) 25 .58
Cases 29(74.4) 10(25.6) 0(0.0) 39
Dral RFLP C/C C/T T/T Sum
R5553668 Total Controls 53(75.7) 15(21.4) 2(2.9) 70 .099a
Cases 75(60.5) 43(34.7) 6(4.8) 124
Boys Controls 32(71.1) 13(28.9) 0(0.0) 45 .25
Cases 58(68.2) 22(25.9) 5(5.9) 85
Girls Controls 2 1 (84.0) 2(8.0) 2(8.0) 25 .001"‘b
Cases l7(43.6) 21(53.8) 1(2.6) 39

 

a: TDT for the allele T of this SNP is significant as p-values of .039 for control vs ADHD-C
(Transmitted: l4, Nontransmitted 5) and .056 for control vs ADHD (C and PI types together) in

total (Transmitted: 22, Nontransmitted 11). In boys, the result is not significant, but, in girls, the

result

is significant, with p-value =.Ol4 for control versus ADHD-C (Transmitted: 6,

Nontransmitted O) & .083 for control vs ADHD-C+PI in total (Transmitted: 9, Nontransmitted 3).
Allele frequency in controls: Mspl (A=.68, G=.33), Hhal (G=.84, A=.l6), Dral (C=.86, T=.l4).

Allele distribution in controls did not differ from Hardy-Weinberg Equilibrium.

64

Discussion
The results suggest that further scrutiny of potential gender specific genetic
correlates of ADHD is necessary. It is unlikely that the associations found in this study
were due to population stratification because ethnicity was relatively well-matched
between cases and controls, and results generally held when analyses were restricted to
Caucasians. However, the small values in each cell lead to a caution that this result needs
to be replicated with larger samples. Each of the three studied genes showed different
patterns of association with ADHD for boys and girls. The insertion polymorphism in
DRD4 was associated with ADHD in boys, while the G allele of the intron 9 SNP in
SLC6A3 and the T allele of the Dral RFLP of the ADRA2A gene were risk alleles in girls.
For SLC6A3 and DRD4, there were no significant differences between boy versus girl
cases or controls, although for SLC6A3 there was a trend toward significance in control
boys versus girls. Especially for the SLC 6A3 intron 9 SNP and VNTR, the genotypes of
control girls are quite different from total controls. This might be due to a protective
effect of common homozygotes rather than a causative effect of rare homozygotes. In
other words, if rare homozygotes are causative and tightly linked to ADHD, then the most
affected children will carry the rare causative allele and most cases would be rare
homozygotes of the rare allele. However, if common homozygotes are highly protective
against ADHD, most controls would be homozygotes for the common allele. The similar
phenomenon is shown in ADRA2A Dral RFLP. Although this effect could depend on the
distribution of population with ADHD symptoms, this is one probable explanation for
these phenomena. For the ADRA2A gene there was a significant difference in allele

frequency when comparing girls versus boys with ADHD.

For DRD4, we found that the main association to ADHD was with the insertion
65

polymorphism, which was also reported by McCracken, et al.13 I. For ADRA2A, the T
allele of the Dral polymorphism was associated with ADHD as we reported previouslyI43 .
However, most previous studies of the dopamine transporter, SLC 6A 3, found that ADHD

35.36

was associated with the 10 repeat allele of the 3’ VNTR‘ . In contrast, our sample

population showed that the 9 repeat version of the VNTR on SLC 6A3 was associated with
ADHD. Swanson, et al. also found the 9 repeat allele was transmitted more frequently to
ADHD children in a sample of methylphenidate responders“. There may be different
subtypes for the 9 or 10 repeat alleles'78"80. For example, differential expression of two
subtypes of the 10 repeat allele has been demonstrated'78"80. As discussed further in the
last chapter, different subtypes of the allele may generate different effects. Other
possibilities include different linkage disequilibrium between a causative allele and the 9
repeat allele in our sample population, or a gender specific effect of the 9 repeat allele
considering the relatively high proportion of girls in our study, or both.

Usually, Bonferroni correction for the multiple testing is too conservative to
apply'gl. In this case, the strict Bonferroni p-value which reaches the .05 significant level
is .002084. Even considering this p-value, the I/D polymorphism of DRD4 in boys and
the DraI RFLP of ADRA2A in girls still remains significant. Regarding this correction,
the significance of SLC6A3 might come out during a multiple testing. There should be an

alpha-inflation correction procedure, but it is still questionable that the strict Bonferroni
correction is appropriate here.

With the present findings that the genetic etiology of ADHD may be different in

girls and boys, there are clear gender differences with respect to the catecholmine system.

In a rat study, the mRNA level and density of the dopamine transporter was significantly

higher in females than in males”. Also, in a human study, SPECT results show the

66

significantly higher density of the dopamine transporter in the striatal region of females“.

For DA management, females have greater striatal DA release and re-uptake than males

'82. A more recognizable result is significant gender and hemisphere

in a rat study
differences in mouse development with overall higher DA level in females'83.

Interestingly, when compared to appropriate controls, ADHD children have higher

dopamine accumulation in the right midbrainm, while ADHD adults have lower

dopamine decarboxylase activity in the medial and left prefrontal areas185 . Moreover,

several studies have shown that ADHD adults have higher dopamine transporter

levels'm‘HZ In ADHD adults methylphenidate treatment initially increases striatal

dopamine transporter activity followed by a reduction in activity afier 4 weeks of

”3. Similar down regulation of the dopamine transporter and the post-synaptic

treatment
dopamine receptor in striatum was seen in ADHD boys after three months of treatment

with methylphenidate”? It is not known how this down regulation occurs, but it is
possible that the dopamine system automatically adjusts to the consistent higher level of
synaptic dopamine caused by methylphenidate treatment. Regulatory control of these
components of the catecholamine system may operate differently in males and females.
ADHD has been described as an inhibition dysfunction. The normal inhibition
process involves prefrontal lobe activation, catecholaminergic transmission from frontal
lobe to striatum, information processing in the striatum, and retransmission from the
striatum. Dysfunction of this prefrontal-striatal network has been implicated in the
etiology of ADHD. Taken together with neuroimaging studies, the suspected etiology of
ADHD is low dopaminergic transmission from the prefrontal lobe to the striatum along
with low striatal activity through high accumulation of dopamine in the striatum, creating

inadequate inhibition through the adrenergic and GABA systems. ADRA2A is closely
67

related to the dopamine system. Clonidine, the ADRA2A agonist sometimes used in
ADHD treatment, is known to reduce dopamine and increase GABA in the nucleus

7
accumbens'8 .

Previous studies suggest that dopamine level is important for proper prefrontal-
striatal function in both girls and boys, but that females may have better systematic

- - - - . 173. 82. 3
management of high levels of dopamine as shown in animal studies I '8

. In managing
larger levels of dopamine in girls, a speculation is that with more active dopamine
transporters, higher synaptic release of dopamine, and probably lower dopamine
receptors in response to released dopamine, a polymorphism that regulates expression of
the dopamine transporter may be more effective in deepening ADHD symptoms than a
polymorphism that regulates expression of the dopamine receptor (Figure 4). Boys may
need more dopamine receptors due to reduced release ability of DA in the synapses, and a
polymorphism that regulates expression of the dopamine receptor would cause more

differences in the receptor level between the boys who have the polymorphism and the

boys who do not have the polymorphism.

 
 

   

\l
<> 0 KMHQDJMK %
, 0* UMUMQU——-~~\_\ 0 ‘ ’ WU ‘ * U “‘9 U U.)
\ /" /

Figure 4. Possible schematic diagram of synapses of girls and boys.

(rectangles: dopamine transporter, small squares: dopamines, and arrows: dopamine receptors)

68

This speculation is consistent with our finding that the dopamine transporter
polymorphism is significantly associated with ADHD in girls rather than boys, whereas
the dopamine receptor polymorphism is more significantly associated with ADHD in
boys than in girls. Also, higher dopamine levels in women may require more or-2A-
adrenergic receptors for adequate regulation of dopamine levels, particularly in the
nucleus accumbens. This agrees with our result that the polymorphism responsible for the
expression of those receptors was more significantly associated with ADHD in girls than

in boys.

Conclusion

Overall, quantitative measurements of inattentiveness and hyperactivity based on
DSM-IV are recommended and gender difference needs to be considered in follow-up
studies. One major concern of the present study is the relatively small sample sizes that
result from splitting the genders or DSM-IV subtypes. Further studies need larger sample

populations to confirm the findings in this chapter.

69

Chapter 4
Haplotype Analysis

Introduction

Based on the initial haplotype analysis of ADRA2A, it is clear that LD between
markers is a very important criterion for marker selection in the candidate gene if the
gene region is in high linkage disequilibrium between polymorphisms. For an example of
two tagging SNPS with allele A (frequency .7) and allele B (frequency .8), if D’ and r2 are
both high, haplotype frequencies would be .7 for AB, .2 for ab, and .1 for aB. If D’ is high
and r2 is low, haplotype frequencies could be .5 for AB, .3 for aB, and .2 for Ab.
Therefore, SNPS with a high value of D’ and low value of r2 are better for capturing the
possible haplotypes in the gene region. In this chapter, three more SNPs farther along the
gene region are selected to find such polymorphisms which may cover additional possible
haplotypes.

The two other candidate genes, SLC6A3 and DRD4 show moderate and low
linkage disequilibrium, respectively, as shown in the literature review in chapter 2. In
examining associations between genes and the disorder, high linkage disequilibrium
between markers may not be necessary“. Bearing in mind the purpose is to find
associations between genes and the disorder, screening all the polymorphisms in
candidate genes would be unnecessarily costly in time and money, therefore the SNPs are
selected based on the spacing and polymorphic status.

In dealing with unambiguous phases in haplotype construction, the program,
PHASE, is used. PHASE implements methods for estimating haplotypes from population

genotype data using a Bayesian statistical methodlxg‘190 Using PHASE, case-control

70

association can be measured in terms of expected frequencies of haplotypes. On the other
hand, the program UNPHASED tests the haplotype-based TDT considering transmission
as a case-control situationm. Reconstructing haplotypes from PHASE may not be highly
reliable in further applications due to the magnification of statistical error; therefore,
case-control tests and QTDTs of individual polymorphisms are conducted for comparison.

Recognizing the better detection through QTDT than TDT, QTDTs are primarily

considered in this chapter.

Updated haplotype analysis for ADRA2A

ADRA2A is located in 10q25.2, near the middle of chromosome 10 and the gene
size is 3,649 bp with one exon. The gene region from 10,000 bp upstream to 2,000 bp
downstream contains a total of 19 SNPs. After screening the gene region for polymorphic
status, three SNPs were typed. Because the previous LD study showed significant D’
within all three typed SNPs, it is not known how far the linkage disequilibrium extends
beyond the gene region. Three more SNPS were selected for the haplotype analysis of
ADRA2A based on the spacing and availability for the assays. The brief summary of
selected SNPs is in Table 16. The first two SNPS, rs638019 and rs491589, are validated
and inventoried assays by Appliedbiosystems (AB1), and the last SNP is non-inventoried

yet functionally tested assay. All markers are in Hardy-Weinberg equilibrium among

founders.

71

Table 16. SNP summary in ADRA2A.

Assay ID c996421 c996423 Mspl Hhal Dral c318157l
rs # rs638019 rs491589 rsl800544 rsl 800545 r5553668 rs602618
Public position 112821869 112824612 1 12826493 1 12827528 1 12829569 1 12833075
distance b/w 2743 1881 1035 2041 3506
Relative position 1 2744 4625 5660 7701 1 1207
allele frequency" .30 .20C .26 .1 1 .19 .42C
allele frequencyID .31 .15 .30 .1 l .18 .31

a: minor allele frequency available from ABI; b: minor allele frequency of all parents; c: minor
allele frequency from NCBI.

As shown in Table 17, the linkage disequilibrium is very high near the ADRA2A
gene region. There is one SNP, Hhal, that shows low r2 with other SNPs. The graphical
summary using GOLD shows this more clearly (Fig 5). Red color represents high LD,
while blue represents low LD. The blue region near Hhal of r2 plot as well as the LD
table indicates that the minor allele of Hhal is usually linked to the minor allele of
c996421, Mspl, and c3181571, but to the major allele of c996423 and Dral. Except for
Hhal, most SNPS are linked to each other through their minor alleles. One interesting
feature of this LD map is that minor alleles of SNPs are sorted with the minor alleles of

either Hhal or c996423 and Dral, which is unexpected if the frequencies of those SNPs

are considered with the allele age.

Table 17. Linkage disequilibrium in ADRA 2A.

D’ (r2)
c996421
c996423

Mspl

Hhal

Dral
c318157l

c996421

c996423
.92 (.35)

Mspl
.93 (.73)
.79 (.30)

72

Hhal

.91 (.19)
.82 (.01)
.89 (.22)

Dral

.95 (.45)
.97 (.77)
.83 (.39)
.92 (.02)

c3181571
.90 (.77)
.89 (.34)
.91 (.75)
.82 (.16)
.91 (.43)

c3181571 -F

Dnla

l-hal -

Mspl -

c996423 —

 

c996421 j-

c3181571-[

Dral —

Hhal -

Mspl -

c996423 ~

 

c996421 J-

iKb 6Kb

l I I ‘1
c996421 Mspl Dral
c996423 Hhal

Figure 5. Graphical summary of LD in ADRA2A.

73

 

c3181571

 

c3181571

 

 

The analyses of individual SNPS are listed below in Table 18 and 19. Similar to
the previous result, the association between ADRA2A and ADHD was more sensitively
detected in QTDT rather than case-control association tests. As expected from previous
results, the girls showed significant associations with ADHD compared to boys. Most
SNPS cannot be tested through QTDT due to smaller samples, but Mspl RFLP and
rs602628 have enough heterozygosity so that some QTDTs are tested. The significant
results for rs602628 suggest that other SNPS, such as rs638019, rs491589, and Dral RF LP,

will probably show more significant results with more girl samples.

Table 18. Case-control association tests for individual SNPS on ADRA2A.

 

 

Assay 1D C996421 C996423 Mspl Hhal Dral c3181571
rs # rs638019 rs491589 r31800544 rsl 800545 r5553668 rs602628
Total dx0 .91 .057 .52 .43 .099 .85
Total dxl .95 .068 .40 .55 .079 .95
Total dx2 .73 .14 .98 .52 .58 .61
Boys dx0 .35 .39 .95 .46 .25 .16
Boys dxl .3 .48 .98 .60 .27 .15
Boys dx2 .91 .17 .81 .51 .18 .70
Girls dx0 .12 .008 .072 .58 .001 .080
Girls dxl .011 .002 .012 .38 .0003 .007
Girls dx2 .88 .10 .71 .93 .035 .81

 

dx0: control vs all ADHD types; dxl: control vs ADHD-C; dx2: control vs ADHD-P1.

Table 19. QTDT for individual SNPS on ADRA2A.

 

 

Assay ID c996421 C996423 Mspl Hhal Dral C318157l
rs # rs638019 RS491589 rsl 800544 rsl 800545 rs553668 rs602628
Total ATTN .0039 .0043 .0991 NS .0282 .0792
Total HYP .0065 .0263 NS NS .0382 NS
Boys ATTN NS .0849 NS NS NS NS
Boys HYP NS NS NS NS NS NS
Girls ATTN NT NT .0599 NT NT .0050
Girls HYP NT NT NT NT NT .0013

 

74

Table 20. Haplotype association tests through PHASE.

 

Samples Indexa haplotype E(freq) E.[Frgcmcontrols)l E[Freq(cases)]

 

 

1 211111 .619 .636 .610

2 122122* .146 .069 .190

Tknal 3 112212 .130 .151 .118

02 4 112122 .037 .065 .021

' 5 111111 .019 .017 .020

6 112112 .014 .025 .008

7 211112 .013 .011 .014

1 211111 .636 .606 .651

2 122122* .136 .078 .166

IBoys 3 112212 .121 .141 .110

17 4 112122 .036 .067 .019

' 5 111111 .027 .036 .022

7 211112 .014 .023 .009

6 112112 .012 .013 .012

1 211111 .578 .670 .519

2 122122* .167 .055 .240

. 3 112212 .137 .157 .125
Chfls

.03 4 112122 .038 .063 .023

7 211112 .020 .006 .029

8 112121 .019 .002 .030

6 112112 .015 .035 .002

 

Estimated haplotype frequencies are listed (E(freq)) and p-values are from 100 permutations. a:

Numbers were marked in the order of the haplotype frequencies from total samples.

Haplotype analysis was done using all the SNPs typed because all SNPs are
linked to each other showing high LD in terms of D’. PHASE gives a p-value of .02 for
testing the significant differences in haplotype frequencies of total samples through
default 100 permutations. The ADHD phenotypes used in this analysis contain both
ADHD-C and ADHD-PI. As expected for the significant association of ADHD girls and
ADRA2A, the p-value was .17 for boys and .03 for girls. The haplotype frequencies which
account for more than 98% of the total possible haplotypes are summarized in Table 20.
As indicated (*), the significance comes mostly from the frequency difference of

haplotype 122122 between controls and cases in all tests: total. boys, and girls. Compared

75

to the individual test result of each SNP, all the alleles in the haplotype 122122 are
associated causatively with ADHD. Although the constitution of SNPS is different, the
frequencies of haplotypes are not much different from the reported haplotype frequencies
also using PHASE.192

UNPHASED gives less significant associations than the association test using
individual SNPs similar to the haplotype analyses using PHASE (Table 21). The results
of individual SNPs using UNPHASED are analogous to the results of QTDT with the
same pattern but different p—values. The haplotype analysis confirms the causative
association between the haplotype 122122 and ADHD. Interestingly, the significant
association with ADHD girls is due to not only the causative effect of the haplotype
122122, but also the protective effect of another haplotype, 211111. The protective

haplotype consists of opposite alleles of the causative haplotype, 122122, except the Hhal

RFLP.

Table 21. Haplotyope association tests through UNPHASED.

 

 

 

 

_ p-value of
Global Assocnated
Samples Tests , b Direction the
p-value“ haplotype
haplotype
T l Inattentiveness .024 1-2-2-1-2-2 Causative .009
ota
Hyperactivity . 13 1-2-2-1-2-2 Causative .026
Inattentiveness .62 1-1-2-1-2-2 - .095
Boys
Hyperactivity .64 1-1-2-1-2-2 - .10
2-1-1-1-1-1 Protective .004
Inattentiveness .004
_ 1-2-2-1-2-2 Causative .016
g . . 2-1-1-1-1-1 Protective .005
H yperactrvrty .012
1-2-2-1-2-2 Causative .007

 

a: permutated 1000 times. b: p-values less than .05 or most associated haplotype.

76

The individual and haplotype association test verified again the association
between the ADRA2A gene and ADHD. Through the individual and haplotype
association, the haplotype 122122 is associated with ADHD in a causative way possibly
due to the contribution of each polymorphism. The protective effect of the haplotype

21 11 11 in the UNPHASED result is possibly due to the protective effect of the haplotype
in girls considering lower p-values than the causative haplotype. Yet, it should be noted
that this most common haplotype, 211111, could be paired often with the causative

haplotype so to make higher significant association between the haplotype 211111 and

ADHD in girls.

Haplotype analysis of SLC6A3

SLC6A3 is located in the telomeric region of chromosome 5 and is much larger in

size than ADRA2A or DRD4. As mentioned in chapter 2, this gene has been previously
examined in association studies and a quite dense LD map is already available. A
previous study on linkage disequilibrium showed two haplotype blocks in the gene region,
one from the promoter to intron 6, the other from exon 9 to the 3’UTR58. In our initial
study, 4 SNPS and a 3’UTR VNTR were typed by PCR, RFLP, or sequencing. The four
SNPS are the exon 9 non-synonymous amino acid change SNP, the intron 9 SNP, and two
exon 15 3’UTR SNPS (Figure 6; indicated with solid arrows). All those polymorphisms
showed relatively high linkage disequilibrium, as expected from the previous study. With
the VNTR as a center, two SNPS of exon 9 and intron 9 and two 3’UTR SNPS showed

higher linkage disequilibrium with respect to their pairwise close locations.

77

5' :
l—H

eve

 

€11

H—1+—+11—1~—+—1—— I

n11

-L .1.

I - coding region I — untranslated region

Figure 6. Gene structure and genotyped polymorphisms of SLC 6A3 .

3.

Arrows from left to right indicate polymorphisms depending on marker numbers shown in table

22.

Table 22. Polymorphism summary in SLC6A3.

 

 

Marker Public Distance Relative Allele Allele
# Assay ID rs # position b/w position freqa freqb
1 C27477615 rs3756450 1501148 9794 l .16 . 15
2 C2960958 rs403636 1491354 5478 9795 .22C .17
3 C3284838 rs465130 1485876 8971 15273 .19 .23
4 C3284822 rs464049 1476905 7763 24244 .49 .44
5 C2396880 rs40358 1469142 4730 32007 . 12 .15
6 Dat1E9 rs6347 1464412 1427 36737 .19 .26
7 Dat119 rs8179029 1462985 14908 38164 09‘" c .16
8 C2960969 rs40184 1448077 958 53072 .39 .44
9 VNTR - 1447119 775 54030 .6883" .76
10 3’UTR SNPl rs3797200 1446344 383 54805 - .22
11 3’UTR SNP2 r51809939 1445961 1592 55188 .21C .22
12 C2854709 - 1444369 6371 56780 .33 .36
13 C2854700 - 1437998 - 63151 .17 .13

 

a: minor allele frequency available from ABl; b: minor allele frequency of parents; c: minor allele

frequency from NCBI; d: various source from publications; e: published reports of the frequency

is similar to the frequency of this sample population“.

Additional SNPs were typed by Taqman genotyping assays. The validated assays

were selected primarily, and further SNPS were selected based on their spacing. The

locations are indicated in Figure 6 with dotted arrows. The information of all the

78

polymorphisms is summarized in Table 22. All markers were checked for Hardy-

Weinberg equilibrium among founders, and markers show no deviation.

Table 23. Linkage disequilibrium in SLC 6A3 .
D’ 2 3 4 5 6 7 8 9 10 11 12 13
.15 .46 .36 .17 .04 .52 .02 .27 .41 .43 .13 .02

y—.

2 - .84 .79 .95 .11 .13 .20 .17 .15 .09 .07 .16
3 - .88 .82 .03 .37 .06 .06 .27 .27 .10 .19
4 — .79 .001 .43 .17 .16 .28 .25 .02 .19
5 - .18 .18 .30 .19 .18 .16 .22 .21
6 - .93 .44 .45 .41 .44 .22 .38
7 - .55 .80 .76 .74 .37 .45
8 - 50 42 .43 03 13
9 - 82 .80 50 51
10 - .88 .47 .48
11 - .41 .52
12 - .45
13 -

As indicated in Table 23 and Figure 7, there are two LD blocks in this gene
region. This LD structure is similar to the LD studies from others which are summarized
in Chapter 2. A further SNP,(marker 1) beyond 3’UTR region was not included in the
first block indicating the sizes of the block. The last SNP (marker 13) shows some LD
with the second blocks, but the LD is decayed due to the SNP, rs40484 (marker 8, ID
c2960969). Considering the high minor allele frequency of the SNP (.44), it is possible
that this SNP is a neutral and old polymorphism showing relatively low LD with nearby

SNPS. Unlike ADRA2A, in which the HapMap project has only one typed SNP, the

79

SNPs. Unlike ADRA2A, in which the HapMap project has only one typed SNP, the
available LD results on the SLC6A3 region were similar for the five SNPS that are typed
in this study and the HapMap. Based on this LD structure, the haplotypes using all

polymorphisms as well as two blocks (markers 2-5 and markers 6-13) were selected for

haplotype analyses.

Figure 7. Graphical summary of LD in SLC6A3 (D’).

 

8K!) 40Kb
L 4 1 I
1 11 11111 |

1 23 4 567 81213

Case—control associations for individual polymorphisms are summarized in Table
24. Most SNPS located in the second LD block are associated with ADHD in the case-

control association tests. Similar to the pattern in chapter 3, girls show more significant
80

with both ADHD-C and ADHD-PI, but c2960969 is associated mostly with ADHD-PI
and the polymorphisms, VNTR, 3’UTR SNPI and 2, c2854709, and C2854700, are

associated mostly with ADHD-C.

Table 24. Case-control association tests for individual polymorphisms on SLC 6A3 .

' Total Total Total Boys Boys Boys Girls Girls Girls
Assay ID

dx0 dxl dx2 dx0 dxl dx2 dx0 dxl dx2

C27477615 .51 .49 .63 .59 .68 .65 .58 .39 .76
C2960958 .65 .47 .86 .31 .087 .20 .88 .67 .63
C3284838 .59 .67 .22 .61 .72 .50 .87 .69 .43
C3284822 .67 .63 .76 .50 .65 .39 .96 .76 .49
C2396880 .70 .70 .83 .42 .17 .16 .93 .61 .66
Dat1E9 .82 .67 .88 .66 .66 .86 .16 .10 .35
Dat119 .046 .060 .11 .74 .72 .71 .005 .007 .004
C2960969 .019 .11 .009 .46 .49 .11 .007 .058 .019
VNTR .25 .32 .32 .94 .83 .25 .01 .002 .39
3’UTR SNP] .11 .17 .25 .78 .96 .17 .028 .011 .23
3’UTR SNP2 .06 .10 .17 .58 .83 .11 .028 .011 .23
C2854709 .10 .033 .53 .33 .28 .91 .049 .019 .083
C2854700 .57 .45 .40 .53 .44 .85 .43 .86 .23

dx0: control vs all ADHD types; dxl: control vs ADHD-C; dx2: control vs ADHD-Pl.

With similar patterns to the results of previous chapters, the QTDT results for
individual SNPs in SLC6A3 show no association between any SNP and ADHD except
rs40184 (ID c2960969). C2960969 showed borderline significance in inattentiveness
with p-value of .035 in total samples. ADHD girls are more significantly associated with

c2960969 (p-values: .0034 for inattentiveness and .030 for hyperactivity).

81

Haplotype analyses using PHASE reaffirm the association between SLC6A3 and
ADHD girls and total samples in the second LD block. Similar to the results in ADRA2A,
the haplotype analysis is not as sensitive as the association tests using individual
polymorphisms (Table 25). The haplotype frequencies consisting of all polymorphisms
are summarized in Table 26 for total samples and girls. The “2” in the haplotypes
represents 10 repeats of VNTR, and “3” represents 9 repeats of VNTR. The haplotypes
from total samples in Table 26 cover only 50 % of all haplotype frequencies with various
haplotypes with the frequency of around .02. Unlike ADRA2A, there is no haplotype

which shows obvious differences in frequencies among listed haplotypes.

Table 25. p-values of the case-control haplotype analyses using PHASE.

 

 

PHASE results All polymorphisms LD block 1 LD block 2
Total .06 .92 .05
Boys .34 .42 .38
Girls .04 .87 .003*

 

Perrnutated 100 times. *: permutated 1000 times.

The results from the second LD block show more obvious differences between
controls and cases (Table 27). The listed haplotypes of total samples cover 85% of all
possible haplotypes. The haplotype 221 3 2212 is associated with ADHD in both total
samples and girls. Minor differences are; a) the most frequent haplotype, l 12 2 1122, is
more frequent in controls suggesting possible protective effect of the haplotype; b) the
haplotype 222 3 2212 is associated with ADHD in total samples, whereas the haplotype
221 3 2222 is associated with ADHD in girls suggesting that the marker Dar/I9 does not

have much effect in total samples, whereas C2854709 may not be effective in girls.

82

Table 26. Summary of haplotype frequencies with all polymorphisms.

 

 

 

Samples
and Index* haplotype E(freq) E[Freq(controls)] E[Freq(cases)]
p-values

l 1212111221122 .131 .149 .121

2 12121112 21112 .063 .051 .070

3 12121112 21121 .063 .044 .073

4 12211112 2 1122 .035 .024 .042

Total 5 12111111 2 1122 .035 .026 .040

06 6 121212213 2212 .034 .027 .038

° 7 1211121121122 .031 .039 .026

8 11122221 3 2222 .030 .018 .036

9 121111113 2212 .029 .025 .031

10 11122112 21122 .026 .031 .024

11 1211111221122 .025 .036 .018

1 12121112 21122 .184 .247 .143

2 12121112 21112 .059 .049 .066

3 12121112 21121 .049 .029 .062

10 11122112 21122 .039 .050 .031

Girls 11 12111112 2 1122 .036 .055 .023

04 12 22211112 2 1122 .034 .040 .030

' 16 121211113 2212 .025 .033 .020

4 12211112 21122 .025 .039 .016

15 12211112 21112 .023 .028 .019

6 121212213 2212 .021 .005 .032

5 1211111121122 .017 .006 .025

 

Pennutated 100 times. *2 Numbers were marked in the order of the haplotype frequencies from

total samples.

83

Table 27. Summary of haplotype frequencies with polymorphisms in the second LD

 

 

 

block.
Samples Indexa haplotype E(freq) E[Freq(controls)] E[Freq(cases)L

1 112 2 1122 .266 .301 .246

2 11221112 .113 .113 .113

3 111 2 1122 .093 .082 .098

4 112 2 1121 .088 .067 .099

Total 5 221 3 2212* .072 .040 .090
.05 6 1 11 3 2212 .064 .073 .059
7 221 3 2222 .051 .038 .057

8 211 2 1122 .048 .060 .041

9 111 2 1121 .038 .030 .043

10 222 3 2212* .021 .003 .031

l 112 2 1122* .327 .462 .240

2 11221112 .138 .132 .143

4 112 2 1121 .077 .064 .085

6 111 3 2212 .055 .053 .056

Girls 5 221 3 2212* .053 .008 .082
.003b 3 111 2 1122 .047 .041 .051
8 211 2 1122 .040 .050 .033

7 221 3 2222* .039 .007 .059

11 11121112 .030 .010 .043

15 11221111 .022 .018 .024

 

Permutated 100 times. a: Numbers were marked in the order of the haplotype frequencies from

total samples; b: permutated 1000 times.

Unexpectedly, the haplotype analyses using UNPHASED reveal an association
between ADHD and SLC6A3. As shown in Table 29, the associated haplotype is 1-2-1-2-
1-1-1-2-2-1-1-2-2, which is the most frequent haplotype in PHASE results. In the second
LD block, the associated haplotype, 1-1-2-2-1-1-1-2, is mostly a part of the haplotype l-
2-1-2-1-1-1-2-2-1-1-2-2 except for the SNP C2854709. An interesting feature is that this
haplotype is protective and the most common haplotype. Moreover, the global p-values
are not significant before permutations possibly suggesting the possible effects of not
only the protective haplotype itself, but also many causative rare haplotypes, which are

paired with this haplotype in each individual. No results of the first LD block show
84

association using either PHASE or UNPHASED.

Table 28. p-values of the association tests from UNPHASED.

 

 

 

 

p-values poljmorphisms ATTN HYP
all .016 .39
Total LD1 .46 .76
LD2 .47 .14
all .5 .016
Boys LD1 1 1
LD2 .25 .033
all .059 .99
Girls LD1 .48 1
LD2 .006 .18

 

Pemiutated 1000 times.

Table 29. Haplotyope association tests through UNPHASED.

 

 

 

 

 

 

 

Global p-value of
Samples Tests p- Associated haplotypeb Direction the
valuea haplotype
All .016 1-2-1-2-1-1-1-2-2-1-1-2-2 protective .017
ATTN
LD2 .47 - - -
Total
All .39 - - -
HYP
LD2 .14 - - -
All .50 - - -
ATTN
LD2 1.00 - - -
Boys
HYP All .016 1-2-1-2-1-1-1-2-2-1-1-2-2 protective .020
LD2 .033 1-1-2-2-1-1-1-2 protective .013
All .059 1-2-1-2-1-1-1-2-2-1-1-2-2 protective .073
ATTN
. LD2 .006 1—1-2-2-1—1-2-2 protective .0155
G1rls
All .99 - - -
HYP
LD2 .18 - - -

 

a: same as Table 24; b: only listed in case ofa significant global p-value.

85

The association tests of individual polymorphisms as well as haplotypes
reconfimi the association between ADHD and SLC 6A3 mostly through the second LD
block. Again, girls show more associations between ADHD and this gene, but the TDT
results for haplotypes of boys show the possible association between hyperactivity and

SLC6A 3 in a protective way.

Haplotype analysis of DRD4

DRD4 is 3398 bp consisting of four exons located in the telomeric region of
chromosome 11. As described in chapter 2, the promoter region of this gene represents a
possible recombination spot in another study, and several evolutionary investigations on
this gene show an LD decay from 7 repeats of VNTR as a center. The SNP selection for
DRD4 is similar to SLC6A3 and ADRA2A. Some of the selected SNPs do not show
enough heterozygosity to be analyzed, and several of them did not work well in the case
of the not-validated but functionally tested assays. Finally, four validated assays and two
functionally tested assays were successfully typed and analyzed. Further typing in the
region between the first two assays and the promoter is not considered because the region
is too far from the gene region. All markers are in Hardy-Weinberg equilibrium among
founders except c7470701, which cannot be tested reliably due to the small values in a

cell because of the low frequency.

 

5' - -- ------ - ————————— I—- 3'
I - coding region I - untranslated reg1on

Figure 8. Gene structure and typed polymorphisms.

86

Table 30. Summary of typed polymorphism in DRD4.

Marker Public Distance Relative Allele Allele
# Assay ID rs # position b/w position freqa freqb
1 Cl611535 - 615085 610 1 .37 .24
2 C1611534 - 615695 10504 611 .26 .32
3 C7470692 rs936460 626199 297 1 1 1 15 .28C .32
4 C7470693 rs93 6461 626496 433 1 1412 .47 .41
5 C7470701 rs916455 626929 915 1 1845 .07c .03
6 In / Del rs4646983 627844 4143 12760 06‘ c .21
7 VNTRf - 631987 2761 16903 ~.19d .20
8 C2565 2468 - 634748 - 19664 .13 .07

a: minor allele frequency available from ABI; b: minor allele frequency of parents calculated by
merlin; c: minor allele frequency from NCBI; d: various source from publications; e: the minor

allele frequency is around .19 in published results'3 I; f: frequency of 7 repeats.

As summarized in Table 31 and Figure 9, the promoter region of this gene
represents relatively low LD between markers which are close. It is suggested that this
low LD region is a recombination hot spot and the polymorphisms in this region are

associated independently from VNTRI22

. Although markers are not enough to interpret,
the LD decays toward the DRD4 gene region, and, after marker 5 (C7470701), the LD
pattern shows inconsistency not depending on the distance. The r2 values are high only

between the first three SNPS, suggesting these SNPs are associated with each other within

their minor alleles.

87

Table 31. Linkage disequilibrium in DRD4.
D’ (12) 2 3 4 5 6 7 8
1 .88 (.58) .57 (.35) .28 (.06) .07(.0001) .09 (.003) .51 .66 (.02)

2 - .86 (.46) .22 (.02) .15 (.002) .74 (.05) .61 .65 (.01)

3 - .30 (.06) .05(.0002) .88 (.09) .43 .23 (.009)
4 - .49 (.01) .59 (.13) .15 .63 (.05)

5 - .48 (.03) .81 1.00(.oo3)
6 - .15 .33 (.03)

7 - .40

Figure 9. Graphical summary of linkage disequilibrium in DRD4 (D’).

7

1111

 

 

 

2Kb 10 Kb

1
II 1111 I I

12 356 7 8

88

Individual polymorphisms were tested for case—control association (Table 32). As

described in the previous chapter, newly typed SNPS show more significant associations

with ADHD boys. Some of the associated polymorphisms show a preference toward a

subtype of ADHD like the results in SLC6A3. Interestingly, the SNP c1611535 is

associated with ADHD especially in ADHD boys, even though this SNP is located far

from the gene region. There are two more genes, interferon regulatory factor 7 (IRF 7)

and mucin and cadherin-like (MUCDHL) near this SNP. More investigation is necessary

for the association of this SNP to figure if other genes or regulatory elements near the

SNP are involved in ADHD, or if a functional polymorphism strongly linked to this SNP

resides in the DRD4 region, although this is not likely considering the low LD with other

polymorphisms in the gene region.

Table 32. The case-control associations of individual polymorphisms.

 

 

Assay ID Total Total Total Boys Boys Boys Girls Girls Girls
dx0 dxl dx2 dx0 dxl dx2 dx0 dxl dx2
C1611535 .005 .029 .038 .003 .009 .088 .59 .76 .45
C1611534 .081 .29 .035 .033 .067 .10 .96 .40 .29
C7470692 .29 .16 .67 .14 .18 .29 .47 .30 .52
C7470693 .082 .075 .39 .047 .067 .063 .37 .40 .63
C7470701 .084 .051 .62 .38 .29 .98 .10 .087 .41
In / Del .003 .003 .20 .002 .002 .35 .14 .13 .44
VNTR .76 .99 .18 .16 .29 .11 .28 .14 .66
C25652468 .49 .41 .91 .75 .52 .48 .49 .73 .41

 

Similar to the results in SLC 6A3, none of the QTDTs finds associations between

ADHD and DRD4, except the association between 4 repeats allele of VNTR and

89

hyperactivity in girls with p-value of .0243. The 4 repeats allele of VNTR is the only one
which shows population stratification in the hyperactivity of girls using QTDT program.
Considering most polymorphisms are not tested due to the small sample size, it may be
possible to find associations through QTDT with an increased sample size.

Haplotype association studies using PHASE are summarized in Table 33. The
haplotypes listed in total samples cover 70% of all haplotypes. Although the p-value
is .008 in total samples, the haplotype frequencies do not differ between cases and
controls suggesting the significance may come from the differences in many haplotypes.
The haplotype that shows most obviously a difference between cases and controls is
11221 1 4 2. Depending on the individual association of polymorphisms, boys are
expected to show more significant results, but both boys and girls do not reveal
associations suggesting the association with total samples may come from many different
haplotypes which are effective only in one gender. The haplotype studies consisting of
polymorphisms only near DRD4 (markers 3-8) show similar patterns but less significant
associations suggesting the significant result of haplotype association comes from the
entire region including first two SNPS (p-values: .07 for total samples; .10 for boys; .44
for girls). Differing from the results of SLC 6A3, none of haplotype results from
UNPHASED shows an association between DRD4 and ADHD as expected from QTDTs
of individual polymorphisms.

The gene DRD4 is associated with ADHD through the case-control associations
in both individual polymorphisms and haplotypes. As described in Chapter 3, the results
here reinforce the possibility that boys are more inclined toward ADHD because of this
gene. Considering the result of total samples in both individual polymorphisms and

haplotype associations, girls may be affected by DRD4, although the effect may be much
90

less than in boys.

Table 33. Haplotype association studies using PHASE.

 

 

 

 

Samples
& Index haplotype E(freq) E[Freq(controls)] E[Freq(cases)]
p-values

1 11111 1 4 2 .326 .253 .367

2 22221 1 7 2 .091 .107 .082

3 22211 1 7 2 .070 .062 .075

4 11121 2 4 2 .050 .064 .043

Total 5 11121 1 4 2 .045 .039 .048

.008” 6 22221 14 2 .036 .033 .037

7 11111 1 7 2 .027 .025 .028

8 11221 1 4 2* .024 .006 .034

9 11111242 .024 .037 .016

10 21121242 .022 .033 .017

1 11111 1 4 2 .344 .260 .387

2 22221 1 7 2 .106 .148 .084

3 22211 1 7 2 .065 .058 .069

4 11121 2 4 2 .042 .046 .039

Boys 6 22221 1 4 2 .031 .029 .032

.06 9 11111 2 4 2 .027 .041 .020

10 21121242 .027 .047 .017

7 11111 1 7 2 .025 .022 .027

11 11111 1 3 2 .025 .000 .038

5 11121 1 4 2 .024 .023 .024

1 11111142 .296 .260 .318

4 11121242 .086 .113 .069

5 11121 1 4 2 .078 .048 .097

3 22211 1 7 2 .068 .056 .076

Girls 7 11111 1 7 2 .054 .044 .060

.48 6 22221 1 4 2 .039 .034 .042

15 21121282 .035 .060 .018

2 22221 1 7 2 .030 .015 .040

8 11221 1 4 2 .025 .003 .040

9 11111242 .023 .033 .017

 

 

Permutated 100 times. *: permutated 1000 times.

91

Conclusion

The association studies using both individual polymorphisms and haplotypes
reveal these candidate genes are associated with ADHD. Although there are some minor
variations depending on the association tests, as shown in chapter 3, ADRA2A and
SLC6A3 are associated in ADHD girls, whereas DRD4 is associated mostly in boys. As
indicated in chapter 2, the TDT and case-control association tests show different results.

Haplotype association tests follow in the same manner to the associations of
individual polymorphisms in terms of both gender differences and differences in the TDT
and case-control associations. However, unlike the expectation in Chapter 1, tests using
haplotypes do not consistently show higher detection of the associations than tests using
individual polymorphisms. In summary, a set of polymorphisms in an LD block shows
better associations and easier interpretation compared to a set of polymorphisms in a low
LD region.

Similar to the results in chapter 2, most of endophenotypes show no significant
associations with p-values less than .05. The associations with individual polymorphisms
are 1) between C2854709 of SLC6A3 and response variability in girls with p-value
of .014; 2) c161 1534 of DRD4 and the stop signal reaction time test in boys with p-value
of .038 through the total evidence of association in the QTDT program. Both SNPS are
associated with ADHD, but there are several other polymorphisms more strongly
associated with ADHD in these gene regions leading to cautions in the endophenotype

studies.

92

Materials and methods
Setting up 384 well plates and quantification

For this haplotype study, one and half 384-well plates are set up. This decision is
due to the fixed scale of ABI assay products. Six 96-well plates are prepared for the
sample preparation of one and half 384-well plates by Biomek 2000 robot. These wells
contain six non-template controls (three for each plate) and three genomic controls (two
for a 384-well plate and one for half 384-well plate). The concentration of samples in
those 96 wells is evaluated by the RNaseP Assay and adjusted. First, the standard curve
for DNA concentration was built using the serial dilution of 10 ng/ul ABI control DNA
(Figure 10). The reaction efficiency was 96.4 % for the ABI DNA and 94.0 % for a
sample DNA in this study.

Three SNPS in ADRA2A were selected based on their spacing and frequencies as
described previouslym. The promoter SNP, rsl80044 (Mspl RFLP), was typed by a
modified amplification of the region using deaza dGTP as described'og. To summarize
briefly, polymerase chain reaction (PCR) was performed in 20 111 reaction mixture
containing 40 ng of genomic DNA, 62.5 11M each dATP, dTTP, dCTP, 31.25 uM dGTP,
31.25 11M deaza dGTP, 1 11M of each primer, 1.5 mM MgC12, 1X PCR buffer, and 0.5
unit of Taq DNA polymerase.

To test the valid concentration range for the Taqman assay, a test assay is
conducted (Figure 10). The blue dots represent the serially diluted ABI control DNA.
Most of them are well clustered as heterozygotes except the lowest concentration which
is 0.35 ng for 5 ul reaction volumn. The scattered pattern follows highly dependent on the
Concentration, for example the highest concentration is located in the coordinate, 2.5 of X

axis and 2.8 of Y axis. The arrow indicates .70 ng of our control DNA, which is lowest

93

concentration after serial dilution. All the other higher concentrations are well clustered
as homozygote of allele X. The highest DNA amount was 45 ng, and all these are
successfully clustered. Depending on the test result, the higher amount of DNA than 1.25
ng with Ct 26.5 is adjusted in setting up the plate for Taqman typing. The 2.25 111 (or 2.5
111) of concentration-adjusted DNA in six 96-well plates were distributed to one and half

384-well plates by Biomek 2000 robot.

Figure 10. Standard curve using serial dilution ofABI control DNA.

ABI control DNA standard curve

33 ~
31 2
29 1

     

Ct

 

 

23 ' y = -3.4121x + 27.498
21 2
R = 0.9975
19 -
17 ‘
1 15 l ______ L 7-1
.1 -o.5 o 0.5 1 1.5

log (nglul)

94

Marker: ITE; 131’ E; :21 '

Allele Y (Mayan-Demand FAN)

Figure 11. Allelic discrimination plot of a test assay.

 

 

3.4

2.9

2.4

1.9

1.4

0.9

fl Call: [Mixed

:J WPIWIQJEJ

Allelic Discrimination Plot

 

 

 

 

1.3 18

MX(MWVIC)

2.8

WWW“ " ‘

X Undotarmhed
t Allele X

9 Both

9 Allele Y

I NTC

 

In the case of limited samples, the whole genome is amplified using the

Genomiphi DNA amplification kit and diluted for distribution to the 96 wells. To

summarize, 1 pl of 20 ng/ul DNA is added to the sample buffer from Amersham

biosciences. The mixture is heated for 3 minutes at 95 °C as a hot start and cooled down

to 4 °C using the ABI 9700 PCR machine. The 9 pl of reaction buffer and 1 pl of the

95

 

enzyme mixture from the Amersham biosciences are added to the cooled mixture, and the
mixture is incubated at 30 °C for 18 hours followed by the deactivation for 10 minutes at
65 °C. The amplified DNAs are diluted by adding 80 pl double-distilled water, and are
stored in -20 °C until usage. The Genomiphi DNA amplification kit successfully
amplifies around 10 pg of DNA from 20 ng DNA.
TaqMan assays

The procedure for the quantification as well as TaqMan genotyping is described
below. The DNA targeted by the probes was amplified in 5 pl reaction mixture containing
1.4 ng or higher amount of DNA, 1X TaqMan universal PCR master mix with No
AmpErase UNG and 1X Assay mix with the recommended thermal cycler condition
consisting of two minute at 50 °C and an initial denaturing step for 10 minutes at 95 °C,
followed by 40 cycles of 15 seconds at 92 °C (for quantification 95 °C) and one minute at
60 °C. The fluorescent probes cut during PCR were read by the ABI Prism 79OOHT
Sequence Detection System using the software, Sequence Detection System version 2.1.
This procedure was mostly conducted at the Michigan State University Genomics
Technology Support Facility.
Sequencing

The farther region of 3’ UTR of SLC6A3 containing SNPS, rs3797200 and
r518009939 was amplified by PCR in 20 pl reaction mixture containing 20 ng of genomic
DNA, 200 pM dNTPs, 1 pM of each primer, 1.5 mM MgC12, 1X PCR buffer, and 0.5
units of Taq DNA polymerase. Amplification included an initial denaturing step at 94 °C
for 3 minutes followed by 35 cycles consisting of 30 seconds at 94 °C, 30 seconds at
60 °C, and one minute at 72 °C, and the final extension step of 5 minutes at 72 °C with

primer sets, 5’-GTGCGTGCCACATCAATAAC-3’ and 5’-AACGAGACAAGGAGGC
96

TGAG-3’. The amount of amplified DNA was measured roughly by comparing to the
standard marker, 100 bp DNA Ladder of New England Biolab (NEB) in 2% agarose gel
stained with ethidium bromide. The 5 pl of amplified DNA was purified using 2 pl of
UAB shrimp alkaline phosphatase (1 units/ pl) and 1 pl of USB exonuclease I (10 units/
pl). The reaction solution was mixed and incubated at 37 °C for 15 minutes followed by
an inactivation at 80 °C for 15 minutes in a thermocycler. The sequencing was performed
with at least 10-40 ng of purified DNA and 30 pmol of sequencing primer (the forward
primer for PCR, 5’-GTGCGTGCCACATCAATAAC-3’) at the Michigan State University
Genomics Technology Support Facility using the ABI Prism 3700 DNA analyzer or ABI
3730 Genetic Analyzer.

Haplotype analysis

Case-control association tests with haplotypes were conducted using a
coalescent-based Bayesian approach implanted in PHASE version 2.1 ”9"”. PHASE can
reconstruct haplotypes from a population constituted of unrelated individuals, and the
case-control association test implemented in PHASE is a permutation-based likelihood
ratio test. The default 100 permutations were performed in most results. If the p-value
was .01 through 100 permutations, 1000 permutations were performed to get a right p-
value.

F or multilocus haplotypes from unphased genotyped data, pedigree transmission
disequilibrium tests were performed using a generalized linear model for quantitative
traits implemented in UNPHASED version 2.40““. The basic frame of this test is similar
to the QTDT, but UNPHASED can handle multilocus haplotypes using EM algorithm.

The results were permutated 1000 times to get the right p-values.

97

Chapter 5

Discussion

Introduction

The catecholamine pathway has been suspected as a main etiology for ADHD
due to pathophysiology.27 Among the genes related to the catecholamine pathway, two
drug target genes, ADRA2A and SLC6A3, as well as DRD4, which has shown the most
reliable association, were selected as candidate genes. The dopamine transporter gene,
SLC6A3, is a target of methylphenidate, a stimulant drug. On the other hand, the a-2A-
adrenergic receptor is a target of non-stimulant medication. As a result of this study, all
three candidate genes, ADRA2A, SLC6A3, and DRD4, show associations with ADHD
through several individual polymorphisms in each gene. Haplotype association studies
also confirm the association between those genes and ADHD, although it is not clear
which polymorphism is functional or how many functional polymorphisms are in each
gene region without consistently higher detection using haplotype analyses.

This study reveals that the candidate gene association approach is sensitive in
detecting the association of relevant genes. However, there are several intriguing findings.
In the present chapter, these are discussed; as is the possible direction of future studies of

complex traits.

LD and functional polymorphisms in haplotype analysis
In current types of haplotype analyses. since not all the polymorphisms in the
gene region can be tested, there is always a possibility for type 11 error by missing

functional polymorphisms, which reside in a gene region. A more elaborate approach is

98

necessary for preventing this possible error. One possibility is to theorize the LD between
the functional polymorphisms and markers, and predict the detection power of

association depending on the likelihood.

ABC lOO-X-Y-ZO/o
ABC lOO-X-Yoo
ABC lOO-X‘Vo ABC ZO/o
ABC 100% AbC Y%

aBC X°/o

Figure 12. One possible description of haplotype structure using three SNPS.

Upper case: common allele; Lower case: rare allele; Most common SNP2A; Least common SNP:

C.

In the previous chapter 2, the importance of polymorphism selection as it
depends on two different LD measurements was discussed. The occurrence of two
markers with high D’ and low r2 is probable when new polymorphisms are generated in a
high LD block. As shown in Figure 13, if a new polymorphism emerged from a
population, with no selection pressure on the SNPs, the haplotype harboring the new
polymorphism, aBC, will be relatively rarer than the original haplotype, ABC. Therefore,
the next polymorphism, AbC, probably also came from the original haplotype, ABC, too.
If there are more ABCs than the other two haplotypes, aBC and AbC, the other
polymorphism, ABC, has more chance to emerge from the original haplotype, ABC.
However, depending on the presence of selection and bottle-neck effects, the more

common haplotype could be different. Also, the chance appearance of a new SNP on a

99

rare haplotype could generate more possibilities. Considering all these possibilities but

focused on only the final haplotype structures, the common haplotypes of a three-marker

model are summarized in Table 34.

 

 

 

Tablei34. Possible haplotypes of three-marker model depending on LD.
LD D’ LD r2 Possible most common haplotypes*
>0.8 All>0.3 ABC>abc>aBC, abC

A&B,C>0.3, B&C<O.1 ABC>abC>ch>aBC

A&B>0.3, C&A,B<0.1 ABC>abC>ABc>aBC

A&C>0.3, B&A,C<0.l ABC>ch>AbC, aBC

B&C>0.3, A&B,C<0.1 ABC>aBC>Abc>AbC

All<0.l ABC>aBC>AbC>ABc
0.4-0.8 All>0.1 ABC>abc>aBC, abC and others

 

A&B,C>0.1, B&C<O.1
B&A,C>O.1, A&C<0.l
C&A,B>0.1, A&B<0.1
A&B>0. l, C&A,B<0.1
A&C>0.l, B&A,C<0.l
B&C>0.1, A&B,C<0.1
All<0.1

ABC>abC>ch>aBC and others
ABC>abC, aBC, Abc and others
ABC>aBC, AbC>ch, Abc and others
ABC>abC>ABc>aBC and others
ABC>aBC>AbC, aBC and others
ABC>aBC>Abc>AbC and others
ABC>aBC>AbC>ABc and others

 

<0.4
L

 

 

ABC>aBC, AbC, ABc, abC, ch, Abc, abc

 

*Allele notation is the same as Figure 12.

Starting from the simplest case, there is one functional polymorphism, which has

LD relationships with markers. If two markers are selected for detecting the functional

polymorphism, the above three-marker model can be applied to predict the association

detection. Likewise, marker model adding one more anonymous marker from the current

selection of markers can be applied to the prediction for one functional polymorphism.

100

 

If there are two functional polymorphisms, the virtual examination using a two-
marker model can be the first step; the construction of a whole-marker model using both
functional polymorphisms and markers is the second step for detecting associations.
Likewise, three or more functional polymorphisms can be calculated. The functional
polymorphisms can be different in their influence on the trait. This can be incorporated
easily if the amount of influence is calculable.

In general, the best detection of a functional polymorphism will involve those
markers with simultaneously high D’ and all low r2 like the last case of D’>.8 in Table 34.
However, in the case of constructing haplotypes with very rare markers, the detection of
causative or protective haplotypes will be difficult due to the low relative frequencies of
those haplotypes compared to the most frequent one. In other words, if three rare alleles
with frequency of .1 are selected for construction of haplotypes, the maximum coverage
has a probability of .7 within all possible haplotypes with all high D’ and all low r2 values.
Therefore, primary allele frequencies for selection criteria would be in the range of .2-.3.
If the LD blocks do not contain such polymorphisms with both high D’ and low r2 values,
highly polymorphic markers, especially synonymous SNPS, can be selected instead to

sort out the possibility of other haplotypes.

Differential detection of associations between TDT and case-control test

In this study, both the case-control association tests and TDTs were performed to
find the association between the candidate genes and ADHD. However, as indicated
previously, these two tests show inconsistency with each other in detecting associations.
The gene ADRA2A reveals an association with ADHD mostly through TDT, whereas

DRD4 and SLC6A3 show associations with ADHD mostly through case-control
101

association tests. In all three genes, separating the data according to the gender usually
gives higher significance than total samples in either TDT or case-control association
tests.

TDT is popular in light of the concern for population stratification in case-control
studies. However, as mentioned earlier, if controls are well matched by ethnicity to
affecteds or the frequencies of marker and disease alleles are not different among ethnic
groups, population stratification is not a problem. The development of TDT starts from
haplotype relative risk (HRR) using family-based controls instead of population-based
controls. The important property of HR regarding the independence of cells in
contingency tables is indicated in recessive casesm, and, from the observation that the
linkage information is only stored in heterozygous parents in the contingency table, TDT
is developed to test a linkage between a disease and marker loci in the presence of
association“. However, it is unclear how TDT is applicable in dominant or additive
cases. It is also noted that TDT and allele sharing statistics are mutually exclusive in
testing a linkage in sib-pairs'gs.

It is well recognized that TDT has much lower power with the same sample size
due to examining the biased transmission from only heterozygous parents. Compared to
HR, which provides a valid test for the association using all parents, TDT tests a linkage
from heterozygous parents in the presence of association. Missing informative
transmission from homozygous parents may lead to a biased result for testing
associations. It is notable that ADRA2A is associated with ADHD girls through the case-
control association tests. Moreover, as mentioned in Chapter 2, the case-control tests
between parents of controls versus parents of cases (ADHD-C) give significant p-values

of .007 for In/ Del polymorphism in DRD4 and .03 for VNTR in SLC6A3. Further
102

investigation is necessary for the explanation of differential detections between TDT and

case-control association tests.

Opposite direction of VNTR on SLC6A3 in our sample population

Although several association studies showed strong association between 3’UTR

VNTR on SLC6A3 and ADHD, the result was not consistent and finally meta-analysis

could not detect si nificance“. More interestin l , the results in this stud as well as
g g Y y

another publication35 find opposite significant association from other results. Also,

several gene expression studies using 3’UTR VNTR showed inconsistency, as well

(Table 35).

Table 35. SLC 6A3 expression assay depending on their VNTR genotypes.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Result Method Cell line Gender Tissue Seq Ref

10>9 Luciferase COS-7 ParentzMale Monkey Kidney Avail ”m

1 0>9 Quantitative Brain - - _ 1%

RT-PCR samples

9>10 GFP SN4741 Male Mouse embryonic - '97
substantia nigra

9> l O Luciferase SK-N-SH Female HS neuroblastoma Avail '70
Hyperdiploid

9>10 Luciferase HEK-293 - HS fetus hypotriploid Avail '78

VS Luciferase SN4741 Male Mouse embryonic - "4
(9>10) substantia nigra

 

 

NS: Not significant.

There are several studies related to the expression difference between genotypes.

Most are performed using VNTR on the 3’UTR region. As shown in Table 31, the results

103

are not consistent depending on the genotypes. Interestingly, Table 32 shows that the
dopamine transporter density using SPECT also replicated this inconsistency. A gene
expression study using haplotypes on SLC 6A3 showed that promoter and intronic variants
affect the transcriptional regulation of SLC 6A3 and suggested that particular
combinations of polymorphisms in haplotypes affect the expression“. This result also
may explain the reason why the dopamine transporter density is not consistent depending
on the genotypes of a single polymorphism although it cannot explain the inconsistency

of expression experiments of VNTR.

Table 36. SC6A3 density depending on their'VNTR genotypes.

 

 

 

 

EesultIMethod Disorder Gender Population Region Ref
NS.* SPECT Schizophrenia 39 controls Striatum '9‘?
['23IlB-CIT 29 patients
10>9 SPECT Alcoholism Avail 12 controls(4F, 8M) Striatum ""’
Total [mlm-CIT 17 patients(5F, 12M)
9> l O SPECT None Avail 30 only controls Striatum 2””
[mini-CIT (13F, 17M)

 

 

 

 

 

 

 

NS: Not significant. *: But, amphetamine-induced decreased in ['23l]lBZM binding potential was

9>10 in each subgroups (controls and schizophrenia).

SPECT results indicate that the dopamine transporter is concentrated in the

striatum, but none of the gene expression studies are performed using cell lines derived
from striatum. It may be possible that different cell lines express the dopamine
transporter gene differentially. However, one study, the gene expression experiment using
quantitative RT—PCR in brain and lymphocyte, suggests the possibility that it may not be

necessary to look at the brain directly, if gene expression pattern is not different

104

 

depending on their genotypes in different individualsm’. It seems that more delicate

research is needed to find out the reason for inconsistency depending on the VNTR
genotypes.

One SPECT study on the dopamine transporter found that the lO-repeat allele
increases in dopamine transporter density'gg, but another group found an opposite result
using the same assayzoo. The other group reported no association between VNTR and
dopamine transporter density'gg. For these SPECT studies, it is possible that there are
many other polymorphisms affecting the gene transcription and translation. Because this
gene is rather large (74,466 bp) and has two distinct block-like structures, it seems
possible that combination of several different functional polymorphisms makes such a
series of inconsistent results. Other polymorphisms in the gene region that affect
expression previously are demonstrated already“. Although their result showed a bit
more expression of the 9-repeat allele than the 10-repeat, it was not significant due to the
existence of more significant polymorphisms via ANOVA.

Obviously, there are inconsistent results in gene expression studies shown in

Table 31. One possible hypothesis is that sequence variability in VNTR may perform a
real role in expression (Figure 13). Using the available sequence, a comparison was done

between the sequences of VNTR and their expression assay results. Three results were

. 17 - 0 . . .
possrble to access the sequences 8 '8 . As shown in Figure 14, four different sequence

combinations were found from those studies.

105

Figure 13. VNTR subtypes.
A: AGGAGCGTGTCCTATCCCCGGACGCATGCAGGGCCCCCAC

B: AGGAGC[A]TGTCCTATCCC [T]GGACGCATGCAGGGCC CCC AC

C: AGGAGCGTGT[A]CTA[C]CCC[A]GGACGCATGCAGGGCCCCCAC (most frequent)
D: AGGAGCGTGT[A]CTA[C]CCC[A]GGA[T]GCATGCAGGGCCCCCAC

E: AGGAGCGTGT[A]CTA[C]CCC[A]G[A]ACGCATGCAGGGCCCCCAC

F: AGGAGCGTGTCCTATCCCCGGAC[CGGAC]GCATGCAGGGCCCCCAC

G: AGGA[A]CGTGT[A]CTA[C]CCC[A]GGA[T]GCATGCAGGGCCCCCAC

H: [T]GGAGCGTGT[AT]TA[C]CCC[A]GGACGCATGCAGGGCCCCCAC

C’: AGGAGCGTGT[A]CTA[C]CCC[A]GGACGCATGCAGGGCCCCCA[T]

C”: [T]GGAGCGTGT[A]CTA[C]CCC[A]GGACGCATGCAGGGCCCCCAC

D’: TGGAGCGTGTACTACCCCAGGATGCATGCAGGGCCCCCAC

Figure 14. SLC 6A3 expression depending on VNTR subtypes.
1) 9 repeat: AABECD’FDC’ strongest

2) 10 repeat: AABECC”FCDC’

3) 9 repeat: AABECD’F CC ’

4) 10 repeat: AABECHFCGC’ weakest

106

 

This is one possibility to explain the inconsistency of associations. As previously
shown, SLC6A3 is associated with ADHD mostly in girls. It is also plausible that there
may be some differential gene expression between gender groups depending on

genotypes, although it should be carefully examined together with the morphological

differences between genders and hormonal differences.

Possibility of genotyping error.

The importance of genotyping error has been recognized even with 1-2 % error
ratezm‘m. Two kinds of errors are indicated; a pedigree error and a genotyping error”.
The easiest detection of this error is to examine Mendelian inheritance of markers.
Among a total of 228 families in this study, several families showed consistent non-
Mendelian inheritance, probably due to one sample-mixed family and three non-
partemity children (two of them have phenotypes). One child without the phenotype
shows non-Mendelian markers twice and many failed genotypings suggesting that this
DNA sample has poor quality. Those individuals are excluded from the data analyses.

Error detection based on the empirical penetrance model during genotyping is
also suggestedzm. In case of a VNTR typing in this study, it is understood that the
genotyped gel can be misread, even by an experienced laboratory technician. After this
correction is once made, no further mistakes were seen, suggesting that proper training
for genotypers can reduce genotyping errors. In the case of high-throughput genotypings
using Taqman, the reading is done automatically through computer software. Due to the
nature of this technology, it is far greater likely that the heterozygotes are true genotypes

than homozygotes. Two such genotyping errors through Mendelian checks are suspected

among ~9500 genotypes suggesting the observed error rate is approximately .0002 for the

107

 

 

TaqMan assays in this study. As indicated in the literature”, there are further genotyping
errors that are consistent with the Mendelian inheritance, and some of which can be

found using additional error checking based on multipoint analyses in the case of cM-

scale linkage studies.

Future studies

The importance of studying complex traits has been increasingly acknowledged,
but outdated concepts and methodologies are still employed. In this transition period, it is
important to recognize again the nature of differences between single gene disorders and
polygenic complex traits. Without handling the polygenic nature in finding relevant genes,

cures for complex disease will be still in a long way from right treatments.

The current linkage studies are moving from traditional Mendelian models to
non-parametric affected sib-pair (ASP) methods. However, ASP methods do not basically
consider the polygenic nature, and so lead to inconsistent results as summarized in
Chapter 1. Apart from the whole genome scans, the candidate gene approaches are
becoming popular due to the study design and many significant findings. This study
reaffirms the sensitivity of candidate gene approaches by finding the associations
between all three candidate genes and ADHD. An important note is this candidate gene
approach also does not handle the polygenic nature of the disease. This flaw may lead to
false positive or negative errors, or inconsistency among different studies, which are hard
to explain.

This study gives three important messages: a) associations cannot be detected

depending on a statistical approach, which means new approach or further explanation

about differential statistical test results are necessary; b) there can be more than one

108

functional polymorphism in a gene region, which can be possibly protective as well as
causative; c) a further study on gender difference on the target trait is necessary.

Considering these findings, more elaborate approaches or new approaches are needed to

find the real nature of complex traits.

109

10.

II.

12.

13.

14.

Literature Cited

Faraone, S. V. & Doyle, A. E. The nature and heritability of attention-
deficit/hyperactivity disorder. Child Adolesc Psychiatr Clin N Am 10, 299-316,
viii-ix (2001).

Wolraich, M. L., Hannah, J. N., Pinnock, T. Y., Baumgaertel, A. & Brown, J.
Comparison of diagnostic criteria for attention-deficit hyperactivity disorder in a
county-wide sample. J Am Acad Child Adolesc Psychiatry 35, 319-24 (1996).

Swanson, J. M. et al. Attention-deficit hyperactivity disorder and hyperkinetic

disorder. Lancet 351, 429-33 (1998).

Scahill, L. & Schwab-Stone, M. Epidemiology of ADHD in school-age children.
Child Adolesc Psychiatr Clin N Am 9, 541-55, vii (2000).

Brown, R. T. et al. Prevalence and assessment of attention-deficit/hyperactivity
disorder in primary care settings. Pediatrics 107, E43 (2001 ).

Fischer, M., Barkley, R. A., Smallish, L. & Fletcher, K. Young adult follow-up of

hyperactive children: self-reported psychiatric disorders, comorbidity, and the role
of childhood conduct problems and teen CD. J Abnorm Child Psychol 30, 463-75

(2002)
Gaub, M. & Carlson, C. L. Gender differences in ADHD: a meta-analysis and
critical review. J Am Acad Child Adolesc Psychiatry 36, 1036-45 (1997).

Faraone, S. V. et al. Genetic heterogeneity in attention-deficit hyperactivity
disorder (ADHD): gender, psychiatric comorbidity, and maternal ADHD. J
Abnorm Psychol 104, 334-45 (1995).

Smalley, S. L. Genetic influences in childhood-onset psychiatric disorders: autism
and attention-deficit/hyperactivity disorder. Am J Hum Genet 60, 1276-82 (1997).

Biederman, J. et al. Further evidence for family-genetic risk factors in attention
deficit hyperactivity disorder. Patterns of comorbidity in probands and relatives
psychiatrically and pediatrically referred samples. Arch Gen Psychiatry 49, 728-

38(1992)

Eaves, L. et al. Genetic and environmental causes of covariation in interview
assessments of disruptive behavior in child and adolescent twins. Behav Genet 30,
321—34 (2000).

Thapar, A., Harrington, R., Ross, K. & McGuffin, P. Does the definition of ADHD
affect heritability? J Am Acad Child Adolesc Psychiatry 39, 1528-36 (2000).

Wasserstein, J. & Lynn, A. Metacognitive remediation in adult ADHD. Treating
executive function deficits via executive functions. Ann N Y Acad Sci 931, 376-84

(2001).
Diagnostic and Statistical Manual of Mental Disorders, 4th edition (DSM-I V)

(American Psychiatric Association, Washington, DC, 1994).

110

15.

16.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

Tannock, R. Attention deficit hyperactivity disorder: advances in cognitive,
neurobiological, and genetic research. J Child Psychol Psychiatry 39, 65-99

(1998).

Castellanos, F. X. & Tannock, R. Neuroscience of attention-deficit/hyperactivity
disorder: the search for endophenotypes. Nat Rev Neurosci 3, 617-28 (2002).
Faraone, S. V. et al. Segregation Analysis of Attention-Deficit Hyperactivity
Disorder. Psychiatric Genetics 2, 257-275 (1992).

Morton, N. E. Trials of segregation analysis by deterministic and macro
simulation (ed. Chakravarti, A.) (Van Nostrand Reinhold, New York, 1984).

Fisher, S. E. et al. A genomewide scan for loci involved in attention-
deficit/hyperactivity disorder. Am J Hum Genet 70, 1183-96 (2002).

Smalley, S. L. et al. Genetic linkage of attention-deficit/hyperactivity disorder on
chromosome l6p13, in a region implicated in autism. Am J Hum Genet 71, 959-
63 (2002).

Ogdie, M. N. et al. A genomewide scan for attention-deficit/hyperactivity disorder
in an extended sample: suggestive linkage on 17p] 1. Am J Hum Genet 72, 1268-
79 (2003).

Ogdie, M. N. et al. Attention deficit hyperactivity disorder: fine mapping supports
linkage to 5pl3, 6q12, l6p13, and 17p11. Am JHum Genet 75, 661-8 (2004).

Bakker, S. C. et al. A whole-genome scan in 164 Dutch sib pairs with attention-
deficit/hyperactivity disorder: suggestive evidence for linkage on chromosomes
7p and 15q. Am JHum Genet 72, 1251-60 (2003).

Arcos-Burgos, M. et al. Attention-deficit/hyperactivity disorder in a population
isolate: linkage to loci at 4q13.2, 5q33.3, llq22, and l7p1 1. Am JHum Genet 75,

998-1014 (2004).
F araone, S. V. Genetics of childhood disorders: XX. ADHD, Part 4: is ADHD

genetically heterogeneous? J Am Acad Child Adolesc Psychiatry 39, 1455-7
(2000).

Kirley, A. et al. Dopaminergic system genes in ADHD: toward a biological
hypothesis. Neuropsychopharmacology 27, 607-19 (2002).

Accardo, P. J ., Blondis, T.A., Whitman, B.Y., & Stein, M.A. Attention Deficits and
Hyperactivity in Children and Adults (Marcel Dekker Inc., 2000).

Pasquale, J. A., Blondis, T. A., Whitman, B. Y. & Stein, M. A. Attention Deficits
and Hyperactivity in Children and Adults (Marcel Dekker, Inc., New York, 2000).
Zhou, Q. Y. & Palmiter, R. D. Dopamine-deficient mice are severely hypoactive,
adipsic, and aphagic. Cell 83, 1197-209 (1995).

Jaber, M., Bloch, B., Caron, M. G. & Giros, B. [Behavioral, cellular and
molecular consequences of the dopamine transporter gene inactivation]. C R

Seances Soc Biol Fil 192, 1127-37 (1998).
Nieoullon, A. Dopamine and the regulation of cognition and attention. Prog

Neurobiol 67, 53-83 (2002).
11 1

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

Solanto, M. V. Dopamine dysfunction in AD/HD: integrating clinical and basic
neuroscience research. Behav Brain Res 130, 65-71 (2002).

Durston, S. A review of the biological bases of ADHD: what have we learned
from imaging studies? Ment Retard Dev Disabil Res Rev 9, 184-95 (2003).

Vaidya, C. J. et al. Selective effects of methylphenidate in attention deficit
hyperactivity disorder: a functional magnetic resonance study. Proc Natl A cad Sci
U SA 95, 14494-9 (1998).

DiMaio, S., Grizenko, N. & Joober, R. Dopamine genes and attention-deficit
hyperactivity disorder: a review. J Psychiatry Neurosci 28, 27-38 (2003).

Cook, E. H., Jr. et al. Association of attention-deficit disorder and the dopamine
transporter gene. Am J Hum Genet 56, 993-8 (1995).

Barr, C. L. et al. Haplotype study of three polymorphisms at the dopamine
transporter locus confirm linkage to attention-deficit/hyperactivity disorder. Biol
Psychiatry 49, 333-9 (2001).

Curran, S. et al. Association study of a dopamine transporter polymorphism and
attention deficit hyperactivity disorder in UK and Turkish samples. Mol
Psychiatry 6, 425-8 (2001).

Daly, G., Hawi, Z., Fitzgerald, M. & Gill, M. Mapping susceptibility loci in

attention deficit hyperactivity disorder: preferential transmission of parental
alleles at DATI, DBH and DRD5 to affected children. Mol Psychiatry 4, 192-6

(1999)

Gill, M., Daly, G., Heron, S., Hawi, Z. & Fitzgerald, M. Confirmation of
association between attention deficit hyperactivity disorder and a dopamine
transporter polymorphism. Mol Psychiatry 2, 311-3 (1997).

Waldman, I. D. et al. Association and linkage of the dopamine transporter gene
and attention-deficit hyperactivity disorder in children: heterogeneity owing to
diagnostic subtype and severity. Am J Hum Genet 63, 1767-76 (1998).

Holmes, J. et al. A family-based and case-control association study of the
dopamine D4 receptor gene and dopamine transporter gene in attention deficit

hyperactivity disorder. Mol Psychiatry 5, 523-30 (2000).

Palmer, C. G. et al. No evidence of linkage or linkage disequilibrium between
DATI and attention deficit hyperactivity disorder in a large sample. Psychiatr
Genet 9, 157-60 (1999).

Roman, T. et al. Attention-deficit hyperactivity disorder: a study of association
with both the dopamine transporter gene and the dopamine D4 receptor gene. Am
J Med Genet 105, 471-8 (2001).

Swanson, J. M. et al. Dopamine genes and ADHD. Neurosci Biolrehav Rev 24, 21-
5 (2000).

Maher, B. S., Marazita, M. L., Ferrell, R. E. & Vanyukov, M. M. Dopamine
system genes and attention deficit hyperactivity disorder: a meta-analysis.
Psychiatr Genet 12, 207-15 (2002).

112

47.

48.

49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

61.

62.

63.

Faraone, S. V., Doyle, A. E., Mick, E. & Biederman, J. Meta-analysis of the
association between the 7-repeat allele of the dopamine D(4) receptor gene and
attention deficit hyperactivity disorder. Am J Psychiatry 158, 1052-7 (2001).
Comings, D. E. et al. Additive effect of three noradrenergic genes (ADRA2a,
ADRAZC, DBH) on attention-deficit hyperactivity disorder and learning
disabilities in Tourette syndrome subjects. Clin Genet 55, 160-72 (1999).

Comings, D. E., Gonzalez, N. S., Cheng Li, S. C. & MacMurray, J. A "line item"
approach to the identification of genes involved in polygenic behavioral disorders:
the adrenergic a1pha2A (ADRA2A) gene. Am J Med Genet 118B, 110-4 (2003).

Roman, T. et al. Is the alpha-2A adrenergic receptor gene (ADRA2A) associated
with attention-deficit/hyperactivity disorder? Am J Med Genet 120B, 116-20

(2003).

Weiss, K. M. & Clark, A. G. Linkage disequilibrium and the mapping of complex
human traits. Trends Genet 18, 19-24 (2002).

Ardlie, K. G., Kruglyak, L. & Seielstad, M. Patterns of linkage disequilibrium in
the human genome. Nat Rev Genet 3, 299-309 (2002).

Gabriel, S. B. et al. The structure of haplotype blocks in the human genome.

Science 296, 2225-2229 (2002).
Phillips, M. S. et al. Chromosome-wide distribution of haplotype blocks and the
role of recombination hot spots. Nature Genetics 33, 382-387 (2003).

Patil, N. et al. Blocks of limited haplotype diversity revealed by high-resolution
scanning of human chromosome 21. Science 294, 1719-1723 (2001).

Zhang, K. et al. Randomly distributed crossovers may generate block-like
pattem's of linkage disequilibrium: an act of genetic drift. Human Genetics 113,
51-59 (2003).

Dawson, E. et al. A first-generation linkage disequilibrium map of human
chromosome 22. Nature 418, 544-548 (2002).

Greenwood, T. A. et al. Segmental linkage disequilibrium within the dopamine
transporter gene. Mol Psychiatry 7, 165-73 (2002).

Greenwood, T. A. et a1. Evidence for linkage disequilibrium between the
dopamine transporter and bipolar disorder. Am J Med Genet 105, 145-51 (2001).

Hawi, Z. et al. Linkage disequilibrium mapping at DATl, DRD5 and DBH
narrows the search for ADHD susceptibility alleles at these loci. Mol Psychiatry 8,
299-308 (2003).

Ueno, S. et al. Identification of a novel polymorphism of the human dopamine
transporter (DATl) gene and the significant association with alcoholism. Mol
Psychiatry 4, 552-7 (1999).

Morino, H. et al. A single nucleotide polymorphism of dopamine transporter gene
is associated with Parkinson's disease. Ann Neurol 47, 528-31 (2000).

Grunhage, F. et al. Systematic screening for DNA sequence variation in the
coding region of the human dopamine transporter gene (DATI ). Mol Psychiatry 5,

113

64.

65.

66.

67.

68.

69.

70.

71.

72.
73.

74.
75.
76.
77.
78.

79.

275-82 (2000).

Greenwood, T. A. & Kelsoe, J. R. Promoter and intronic variants affect the
transcriptional regulation of the human dopamine transporter gene. Genomics 82,
511-20 (2003).

Cantwell, D. P. & Baker, L. Attention deficit disorder with and without
hyperactivity: a review and comparison of matched groups. J Am Acad Child
Adolesc Psychiatry 31, 432-8 (1992).

Lahey, B. B. & Carlson, C. L. Validity of the diagnostic category of attention
deficit disorder without hyperactivity: a review of the literature. J Learn Disabil
24, 110-20 (1991).

Todd, R. D. et al. Discrimination of DSM-IV and latent class attention-
deficit/hyperactivity disorder subtypes by educational and cognitive performance
in a population-based sample of child and adolescent twins. J Am Acad Child
Adolesc Psychiatry 41, 820-8 (2002).

Todd, R. D. et al. Familiality and heritability of subtypes of attention deficit
hyperactivity disorder in a population sample of adolescent female twins. Am J
Psychiatry 158, 1891-8 (2001).

Gottesman, II & Gould, T. D. The endophenotype concept in psychiatry:
etymology and strategic intentions. Am J Psychiatry 160, 636-45 (2003).
Gottesman, M. M. Helping parents make sense of ADHD diagnosis and treatment.

J Pediatr Health Care 17, I49-53 (2003).

Nigg, J. T., Butler, K. M., Huang-Pollock, C. L. & Henderson, J. M. Inhibitory
processes in adults with persistent childhood onset ADHD. J Consult Clin Psychol
70, 153-7 (2002).

Nigg, J. T. Is ADHD a disinhibitory disorder? Psychol Bull 127, 571-98 (2001).
Sergeant, J. A., Oosterlaan, J ., and van der Meere, J. in Handbook of Disruptive
Behavior Disorders (ed. Quay, H. C. a. H., A.E.) 75-104 (Kluwer Academic/
Plenum Publishers, New York, 1999).

Gershon, J. A meta-analytic review of gender differences in ADHD. J A tten
Disord 5, 143-54 (2002).

Rhee, S. R., Waldman, I.D., Hay, D.A., & Levy, F. in Attention, genes, and ADHD
(ed. Hay, F. L. a. D. A.) 139-156 (Brunner-Routledge, Philadelphia, 2001).

Stoll, M. et al. A genomic-systems biology map for cardiovascular function.
Science 294, 1723-6 (2001).

Moreno, C. et al. Genomic map of cardiovascular phenotypes ofhypertension in
female Dahl S rats. Physiol Genomics 15, 243-57 (2003).

Rivest, R., Falardeau, P. & Di Paolo, T. Brain dopamine transporter: gender
differences and effect of chronic haloperidol. Brain Res 692, 269-72 (1995).
Morissette, M., Biron, D. & Di Paolo, T. Effect of estradiol and progesterone on
rat striatal dopamine uptake sites. Brain Res Bull 25, 419-22 (1990).

114

 

80.

81.

82.

83.

84.

85.

86.

87.

88.

89.

90.

91.

92.

93.

94.

Morissette, M. & Di Paolo, T. Effect of chronic estradiol and progesterone
treatments of ovariectomized rats on brain dopamine uptake sites. J Neurochern
60, 1876-83 (1993).

Zhou, W., Cunningham, K. A. & Thomas, M. L. Estrogen regulation of gene
expression in the brain: a possible mechanism altering the response to
psychostimulants in female rats. Brain Res Mol Brain Res 100, 75-83 (2002).
Bosse, R., Rivest, R. & Di Paolo, T. Ovariectomy and estradiol treatment affect
the dopamine transporter and its gene expression in the rat brain. Brain Res Mol
Brain Res 46, 343-6 (1997).

Staley, J. K. et al. Sex differences in [123I]beta-CIT SPECT measures of
dopamine and serotonin transporter availability in healthy smokers and
nonsmokers. Synapse 41, 275-84 (2001).

Evans, J. M. et a1. Gender differences in autonomic cardiovascular regulation:
spectral, hormonal, and hemodynamic indexes. J Appl Physiol 91, 2611-8 (2001).

Yonezawa, A. et al. Alpha 2-adrenoceptor antagonists: effects on ejaculation,
penile erection and pelvic thrusting behavior in dogs. Pharmacol Biochem Behav
70, 141-7 (2001).

Chen, D. C., Duckles, S. P. & Krause, D. N. Postjunctional alpha2-adrenoceptors
in the rat tail artery: effect of sex and castration. Eur J Pharmacol 372, 247-52
(1999)

Greenhill, L. in Stimulant drugs and ADHD: Basic and Clinical Neuroscience (ed.
M.V. Solanto, A. F. T. A., & F.X. Castellanos) 31-72 (Oxford University Press,
New York, 2001).

Lowe, N. et a]. Joint analysis of the DRD5 marker concludes association with
attention-deficit/hyperactivity disorder confined to the predominantly inattentive
and combined subtypes. Am J Hum Genet 74, 348-56 (2004).

Amsten, A. F., Steere, J. C. & Hunt, R. D. The contribution of alpha 2-
noradrenergic mechanisms of prefrontal cortical cognitive function. Potential
significance for attention-deficit hyperactivity disorder. Arch Gen Psychiatry 53,
448-55 (1996).

Amsten, A. F. T. in Stimulant drugs and ADHD: Basic and Clinical Neuroscience
(ed. M.V. Solanto, A. F. T. A., & F.X. Castellanos) 185-208 (Oxford University
Press, New York, 2001).

Barridge, C. W. in Stimulant drugs and ADHD: Basic and Clinical Neuroscience
(ed. M.V. Solanto, A. F. T. A., & F.X. Castellanos) 158-184 (Oxford University
Press, New York, 2001).

Posner, M. P., S. The attention system of the human brain. Annual Review of
Neuroscience 13, 25-42 (1990).

Berger, A. & Posner, M. I. Pathologies of brain attentional networks. Neurosci
Biobehav Rev 24, 3-5 (2000).

Sergeant, J. A., Oosterlaan, J ., & can der Meere, J. in Handbook ofdisruptive

115

 

95.

96.

97.

98.

99.

100.

101.

102.

103.

104.

105.

106.

107.

108.

109.

behavior disorders (ed. Hogan, H. C. Q. A. E.) 75-104 (Kluwer/Plenum, New
York, 1999).

van der Meere, J. & Stemerdink, N. The development of state regulation in
normal children: An indirect comparison with children with ADHD.
Developmental Neuropsychology 16, 213-225 (1999).

Barkley, R. A. Behavioral inhibition, sustained attention, and executive functions:
constructing a unifying theory of ADHD. Psychol Bull 121, 65-94 (1997).

Shaffer, D., Fisher, P., & Lucas. NIMH Diagnostic Interview Schedule for
Children-J V (Columbia University, New York: Ruane Center for Early Diagnosis,
Division of Child Psychiatry, 1997).

Zentall, S. 2., T. Optimal stimulation: A model of disordered activity and
performance in normal and deviant children. Psychol Bull 94, 446-471 (1983).

McCracken, J. T. A two-part model of stimulant action on attention-deficit
hyperactivity disorder in children. J Neuropsychiatry Clin Neurosci 3, 201-9
(1991)

Malone, M. A., Kershner, J. R. & Swanson, J. M. Hemispheric Processing and
Methylphenidate Effects in Attention-Deficit Hyperactivity Disorder. Journal of
Child Neurology 9, 181-189 (1994).

Barry, R. J ., Johnstone, S.'J. & Clarke, A. R. A review of electrophysiology in
attention-deficit/hyperactivity disorder: 11. Event-related potentials. Clin
Neurophysiol 114, 184-98 (2003).

Losier, B. J., McGrath, P. J. & Klein, R. M. Error patterns on the continuous
performance test in non-medicated and medicated samples of children with and
without ADHD: a meta-analytic review. J Child Psychol Psychiatry 37, 971-87
(1996)

Huang-Pollock, C. L. & Nigg, J. T. Searching for the attention deficit in attention
deficit hyperactivity disorder: The case of visuospatial orienting. C [in Psychol
Rev 23, 801-30 (2003).

Oosterlaan, J., Logan, G D. & Sergeant, J. A. Response inhibition in AD/HD, CD,
comorbid AD/HD + CD, anxious, and control children: a meta-analysis of studies
with the stop task. J Child Psychol Psychiatry 39, 411-25 (1998).

Jakala, P. et a1. Guanfacine, but not clonidine, improves planning and working
memory performance in humans. Neuropsychopharmacology 20, 460-70 (1999).

Biederman, J. & Spencer, T. Attention-deficit/hyperactivity disorder (ADHD) as a
noradrenergic disorder. Biol Psychiatry 46, 1234-42 (1999).

Franowicz, J. S. & Amsten, A. F. The alpha-2a noradrenergic agonist, guanfacine,
improves delayed response performance in young adult rhesus monkeys.
Psychopharmacology (Berl) 136, 8-14 (1998).

Tims, F. M. et al. Characteristics and problems of 600 adolescent cannabis abusers
in outpatient treatment. Addiction 97 Suppl 1, 46-57 (2002).

Xu, C. et al. Linkage study of the alpha2A adrenergic receptor in attention-deficit

116

110.

111.

112.

113.

114.

115.

116.

117.

118.

119.

120.

121.

122.

123.

124.

125.

hyperactivity disorder families. Am J Med Genet 105, 159-62 (2001).

Krause, K. H., Dresel, S. H., Krause, J., 1a Fougere, C. & Ackenheil, M. The
dopamine transporter and neuroimaging in attention deficit hyperactivity disorder.
Neurosci Biobehav Rev 27, 605-13 (2003).

Dougherty, D. D. et al. D0pamine transporter density in patients with attention
deficit hyperactivity disorder. Lancet 354, 2132-3 (1999).

Madras, B. K., Miller, G. M. & Fischman, A. J. The dopamine transporter:
relevance to attention deficit hyperactivity disorder (ADHD). Behav Brain Res
130, 57-63 (2002).

Krause, K. H., Dresel, S. H., Krause, J., Kung, H. F. & Tatsch, K. Increased
striatal dopamine transporter in adult patients with attention deficit hyperactivity
disorder: effects of methylphenidate as measured by single photon emission
computed tomography. Neurosci Lett 285, 107-10 (2000).

Uhl, G. R. & Lin, Z. The top 20 dopamine transporter mutants: structure-function
relationships and cocaine actions. EurJ Pharmacol 479, 71-82 (2003).

Xie, G. X., Jones, K., Peroutka, S. J. & Palmer, P. P. Detection of mRNAs and
alternatively spliced transcripts of dopamine receptors in rat peripheral sensory
and sympathetic ganglia. Brain Res 785, 129-35 (1998).

Edvardsen, O. & Dahl, S. G. A putative model of the dopamine transporter. Brain
Res Mol Brain Res 27, 265-74 (1994).

Loland, C. J ., Norgaard-Nielsen, K. & Gether, U. Probing dopamine transporter
structure and function by Zn(2+)-site engineering. Eur J Pharmacol 479, 187-97
(2003).

Cichon, S., Nothen, M. M., Wolf, H. K. & Propping, P. Lack of imprinting of the
human dopamine D4 receptor (DRD4) gene. Am J Med Genet 67, 229-31 (1996).

Heiser, P. et al. Molecular genetic aspects of attention-deficit/hyperactivity
disorder. Neurosci Biobehav Rev 28, 625-41 (2004).

Grady, D. L. et a1. High prevalence of rare dopamine receptor D4 alleles in
children diagnosed with attention-deficit hyperactivity disorder. Mol Psychiatry 8,
536-45 (2003).

Leung, P. W. et al. Dopamine receptor D4 (DRD4) gene in Han Chinese children
with attention-deficit/hyperactivity disorder (ADHD): increased prevalence of the
2-repeat allele. Am J Med Genet B Neuropsychiatr Genet 133, 54-6 (2005).

Lowe, N. et al. Multiple marker analysis at the promoter region of the DRD4 gene
and ADHD: evidence of linkage and association with the SNP -616. Am J Med
Genet B Neuropsychiatr Genet 13], 33-7 (2004).

Ding, Y. C. et al. Evidence of positive selection acting at the human dopamine
receptor D4 gene locus. Proc Natl Acad Sci U S A 99, 309-14 (2002).

Wang, E. et al. The genetic architecture of selection at the human dopamine
receptor D4 (DRD4) gene locus. Am J Hum Genet 74, 931-44 (2004).

Meulenbelt, 1., Droog, S., Trommelen, G. J., Boomsma, D. I. & Slagboom, P. E.
117

126.

127.

128.

129.
130.

131.

132.

133.

134.

135.

136.

137.

138.

139.

140.

141.

142.

High-yield noninvasive human genomic DNA isolation method for genetic studies
in geographically dispersed families and populations. Am J Hum Genet 57, 1252-
4(1995)

Wechsler, D. Wechsler Intelligence Scale for Children (3rd ed.) (Psychological
Corporation, San Antonio, TX, 1991).

Reynolds, C. K., R. Behavior Assessment System/or Children: Manual.
(American Guidance Service, Inc, MN, 1992).

Conners, K. Conners'Rating Scales-Revised Technical Manual (Multi Health
Systems, NY, 1997).

(ed. Association, A. P.) (APA Press, Washington, DC, 2000).

Lahey, B. B. et al. DSM-IV field trials for attention deficit hyperactivity disorder
in children and adolescents. Am J Psychiatry 151, 1673-85 (1994).

McCracken, J. T. et a1. Evidence for linkage of a tandem duplication
polymorphism upstream of the dopamine D4 receptor gene (DRD4) with attention
deficit hyperactivity disorder (ADHD). Mol Psychiatry 5, 531-6 (2000).

Kang, A. M., Pakstis, A. J. & Kidd, K. K. Linkage disequilibrium at the dopamine
transporter locus (SLC6A3). American Journal of Human Genetics 71, 449-449
(2002).

Willcutt, E. G., Pennington, B. F. & DeFries, J. C. Etiology of inattention and
hyperactivity/impulsivity in a community sample of twins with learning

difficulties. J Abnorm Child Psychol 28, 149-59 (2000).

Wolfarth, B. et al. A polymorphism in the alpha2a-adrenoceptor gene and
endurance athlete status. Med Sci Sports Exerc 32, 1709-12 (2000).

Lario, S. et al. Mspl identifies a biallelic polymorphism in the promoter region of
the alpha 2A-adrenergic receptor gene. Clin Genet 51, 129-30 (1997).

Feng, J. et a1. Variants in the a1pha2A AR adrenergic receptor gene in psychiatric
patients. Am JMed Genet 81, 405-10 (1998).

Hoehe, M. R., Berrettini, W. H. & Lentes, K. U. Dra I identifies a two allele DNA
polymorphism in the human alpha 2-adrenergic receptor gene (ADRAR), using a
5.5 kb probe (p ADRAR). Nucleic Acids Res 16, 9070 (1988).

Abecasis, G. R. & Cookson, W. O. GOLD--graphica1 overview of linkage
disequilibrium. Bioinformatics 16, 182-3 (2000).

Pritchard, J. K. & Przeworski, M. Linkage disequilibrium in humans: models and
data. Am JHum Genet 69, 1-14 (2001).

Comings, D. E. Clinical and molecular genetics of ADHD and Tourette syndrome.
Two related polygenic disorders. Ann N Y Acad Sci 931, 50-83 (2001).

Abecasis, G. R., Cardon, L. R. & Cookson, W. O. A general test of association for
quantitative traits in nuclear families. Am J Hum Genet 66, 279-92 (2000).

Abecasis, G R., Cookson, W. O. & Cardon, L. R. Pedigree tests of transmission
disequilibrium. EurJ Hum Genet 8, 545-51 (2000).

118

143.

144.

145.

146.

147.

148.

149.

150.

151.

152.

153.

154.

155.

156.

157.

158.

Park, L. et a1. Association and linkage of alpha-2A adrenergic receptor gene
polymorphisms with childhood ADHD. Mol Psychiatry (2004).

Ewens, W. J. & Spielman, R. S. The transmission/disequilibrium test: history,
subdivision, and admixture. Am J Hum Genet 57, 455-64 ( 1995).

Milich, R., Balentine, A. C. & Lynam, D. R. ADHD combined type and ADHD
predominantly inattentive type are distinct and unrelated disorders. Clinical
Psychology-Science and Practice 8, 463-488 (2001).

Barkley, R. A., Grodzinsky, G & DuPaul, G. J. Frontal lobe functions in attention
deficit disorder with and without hyperactivity: a review and research report. J
Abnorm Child Psychol 20, 163-88 (1992).

Doyle, A. E., Biederman, J ., Seidman, L. J ., Weber, W. & Faraone, S. V.
Diagnostic efficiency of neuropsychological test scores for discriminating boys

with and without attention deficit-hyperactivity disorder. J Consult Clin Psychol
68, 477-88 (2000).

Pennington, B. F. & Ozonoff, S. Executive firnctions and developmental
psychopathology. J Child Psychol Psychiatry 37, 51-87 (1996).

Jensen, P. S., Martin, D. & Cantwell, D. P. Comorbidity in ADHD: implications
for research, practice, and DSM-V. J Am Acad Child Adolesc Psychiatry 36, 1065-
79(1997)

Hinshaw, S. in Attention deficit hyperactivity disorder: State of the Science, Best
Practices. (ed. Jensen P, C. J.) 1-16 (Civic Research Institute, Kingston, NJ, 2002).

Faraone, S. V., Biederman, J. & Friedman, D. Validity of DSM-IV subtypes of
attention-dcficit/hyperactivity disorder: a family study perspective. J Am Acad
Child Adolesc Psychiatry 39, 300-7 (2000).

Kelleher, K. in Attention deficit hyperactivity disorder: State of the Science, Best
Practices. (ed. Jensen P, C. J.) 1-12 (Civic Research Institute, Kingston, NJ, 2002).

Rietveld, M. J., Hudziak, J. J ., Bartels, M., van Beijsterveldt, C. E. & Boomsma,
D. I. Heritability of attention problems in children: longitudinal results from a
study of twins, age 3 to 12. J Child Psychol Psychiatry 45, 577-88 (2004).

Faraone, S. V., Perlis, R.H., Doyle, A.E., Smoller, J.W., Goralnick, J .J., Holmgren,
M.A., & Sklar, P. Molecular genetics of attention-deficit hyperactivity disorder.
Biologicva] Psychiatry in press (2004).

Almasy, L. & Blangero, J. Endophenotypes as quantitative risk factors for
psychiatric disease: rationale and study design. Am J Med Genet 105, 42-4 (2001).

Giedd, J. N., Blumenthal, J., Molloy, E. & Castellanos, F. X. Brain imaging of
attention deficit/hyperactivity disorder. Ann N Y Acad Sci 931, 33-49 (2001).

Logan, G. D., Schachar, R. J. & Tannock, R. Impulsivity and inhibitory control.
Psychological Science 8, 60-64 (1997).

Nigg, J. T. The ADHD response-inhibition deficit as measured by the stop task:
replication with DSM-IV combined type, extension, and qualification. J Abnorm
Child Psychol 27, 393-402 (1999).

119

159.

160.

161.

162.

163.

164.

165.

166.

167.

168.

169.

170.

171.

172.

173.

174.

175.

Band, G. P., van der Molen, M. W. & Logan, G. D. Horse-race model simulations
of the stop-signal procedure. Acta Psycho] (Amst) 112, 105-42 (2003).

Wolraich, M. L., Hannah, J. N., Baumgaertel, A. & Feurer, I. D. Examination of
DSM-IV criteria for attention deficit/hyperactivity disorder in a county-wide
sample. J Dev Behav Pediatr 19, 162-8 (1998).

Taylor, E. W. & Keltner, N. L. Messy purse girls: Adult females and ADHD.
Perspectives in Psychiatric Care 38, 69-72 (2002).

Rhee, S. H. & Waldman, I. D. Etiology of sex differences in the prevalence of
ADHD: an examination of inattention and hyperactivity-impulsivity. Am J Med
Genet 1278, 60-4 (2004).

Ernst, M. et al. Reduced brain metabolism in hyperactive girls. J Am Acad Child
Adolesc Psychiatry 33, 858-68 (1994).

Nigg, J. T., Blaskey, L. G., Huang-Pollock, C. L. & Rappley, M. D.
Neuropsychological executive functions and DSM-1V ADHD subtypes. J Am
Acad Child Adolesc Psychiatry 41, 59-66 (2002).

Giedd, J. N., Castellanos, F. X., Rajapakse, J. C., Vaituzis, A. C. & Rapoport, J. L.
Sexual dimorphism of the developing human brain. Prog Neuropsychopharmaco]
Biol Psychiatry 21, 1185-201 (1997).

Breedlove, S. M. Sexual differentiation of the human nervous system. Annu Rev
Psychol 45, 389-418 (1994).

Hafner, H. Gender differences in schizophrenia. Psychoneuroendocrinology 28
Suppl 2, 17-54 (2003).

Seeman, M. V. Psychopathology in women and men: focus on female hormones.
Am JPsychiatry 154, 1641-7 (1997).

Biederman, J. et a1. Gender differences in a sample of adults with attention deficit
hyperactivity disorder. Psychiatry Res 53, 13-29 (1994).

Breedlove, S. M. a. H., E. in Behavioral Endocrinology (ed. al., B. e.) 225-263
(2002).

Takagi, G. et al. Gender differences on the effects of aging on cardiac and
peripheral adrenergic stimulation in old conscious monkeys. Am J Physiol Heart
Circ Physiol 285, H527-34 (2003).

Ghali, J. K. et al. Gender differences in advanced heart failure: insights from the
BEST study. J Am Coll Cardiol 42, 2128-34 (2003).

Festa, E. D. et al. Sex differences in cocaine-induced behavioral responses,
pharmacokinetics, and monoamine levels. Neuropharmacology 46, 672-687
(2004).

Dluzen, D. E., Tweed, C., Anderson, L. I. & Laping, N. J. Gender differences in
methamphetamine-induced mRNA associated with neurodegeneration in the

mouse nigrostriatal dopaminergic system. Neuroendocrinology 77, 232-238
(2003).

Laakso, A. et al. Sex differences in striatal presynaptic dopamine synthesis
120

176.

177.

178.

179.

180.

181.

182.

183.

184.

185.

186.

187.

188.

189.

190.

capacity in healthy subjects. Biol Psychiatry 52, 759-63 (2002).

Wong, D. F. et al. Invivo Measurement of Dopamine-Receptors in Human-Brain
by Positron Emission Tomography - Age and Sex-Differences. Annals of the New
York Academy of Sciences 515, 203-214 (1988).

Andersen, S. L. & Teicher, M. H. Sex differences in dopamine receptors and their
relevance to ADHD. Neuroscience and Biobehavioral Reviews 24, 137-141
(2000).

Miller, G. M. & Madras, B. K. Polymorphisms in the 3'-untranslated region of
human and monkey dopamine transporter genes affect reporter gene expression.
Mol Psychiatry 7, 44-55 (2002).

Inoue-Murayama, M. et a1. Variation of variable number of tandem repeat
sequences in the 3'-untranslated region of primate dopamine transporter genes that
affects reporter gene expression. Neurosci Lett 334, 206-10 (2002).

Fuke, S. et al. The VNTR polymorphism of the human dopamine transporter
(DATl) gene affects gene expression. Pharmacogenomics J 1, 152-6 (2001).

Garcia, L. V. Escaping the Bonferroni iron claw in ecological studies. Oikos 105,
657-663 (2004).

Walker, O. D., Rooney, M. B., Wightman, R. M. & Kuhn, C. M. Dopamine
release and uptake are greater in female than male rat striatum as measured by fast
cyclic voltammetry. Neuroscience 95, 1061-70 (2000).

Connell, S., Karikari, C. & Hohmann, C. F. Sex-specific development of cortical
monoamine levels in mouse. Brain Res Dev Brain Res 151, 187-91 (2004).

Ernst, M. et a1. High midbrain [18F ]DOPA accumulation in children with
attention deficit hyperactivity disorder. Am J Psychiatry 156, 1209-15 (1999).

Ernst, M., Zametkin, A. J ., Matochik, J. A., Jons, P. H. & Cohen, R. M. DOPA
decarboxylase activity in attention deficit hyperactivity disorder adults. A

[fiuorine- l 8]f1uorodopa positron emission tomographic study. J Neurosci 18,
5901-7 (1998).

Vles, J. S. et a1. Methylphenidate down-regulates the dopamine receptor and
transporter system in children with attention deficit hyperkinetic disorder
(ADHD). Neuropediatrics 34, 77-80 (2003).

Murai, T. et al. Clonidine reduces dopamine and increases GABA in the nucleus
accumbens: an in vivo microdialysis study. Pharmacol Biochem Behav 60, 695-
701 (1998).

Cardon, L. R. & Abecasis, G R. Using haplotype blocks to map human complex
trait loci. Trends Genet 19, 135-40 (2003).

Stephens, M., Smith, N. J. & Donnelly, P. A new statistical method for haplotype
reconstruction from population data. Am J Hum Genet 68, 978-89 (2001).

Stephens, M. & Donnelly, P. A comparison of bayesian methods for haplotype
reconstruction from population genotype data. Am J Hum Genet 73, 1162-9
(2003).

121

191.

192.

193.

194.

195.

196.

197.

198.

199.

200.

201.

202.

203.

204.

205.

Dudbridge, F. Pedigree disequilibrium tests for multilocus haplotypes. Genet
Epidemiol 25, 115-21 (2003).

Belfer, I. et al. Haplotype-based analysis of alpha 2A, 28, and 2C adrenergic
receptor genes captures information on common functional loci at each gene. J
Hum Genet 50, 12-20 (2005).

Stephens, M. & Donnelly, P. A comparison of bayesian methods for haplotype
reconstruction from population genotype data. Am J Hum Genet 73, 1162-9
(2003).

Ott, J. Statistical properties of the haplotype relative risk. Genet Epidernio] 6, 127-
30 (1989).

Spielman, R. S. & Ewens, W. J. A sibship test for linkage in the presence of
association: the sib transmission/disequilibrium test. Am J Hum Genet 62, 450-8
(1998).

Mill, J ., Asherson, P., Browes, C., D'Souza, U. & Craig, 1. Expression of the
dopamine transporter gene is regulated by the 3' UTR VNTR: Evidence from
brain and lymphocytes using quantitative RT-PCR. Am J Med Genet 114, 975-9
(2002).

Michelhaugh, S. K., Fiskerstrand, C., Lovejoy, E., Bannon, M. J. & Quinn, J. P.
The dopamine transporter gene (SLC6A3) variable number of tandem repeats
domain enhances transcription in dopamine neurons. J Neurochem 79, 1033-8
(2001).

Martinez, D. et al. The variable number of tandem repeats polymorphism of the
dopamine transporter gene is not associated with significant change in dopamine
transporter phenotype in humans. Neuropsychopharmacology 24, 553-60 (2001).

Heinz, A. et al. Genotype influences in vivo dopamine transporter availability in
human striatum. Neuropsychopharmacology 22, 133-139 (2000).

Jacobsen, L. K. et al. Prediction of dopamine transporter binding availability by
genotype: a preliminary report. Am J Psychiatry 157, 1700-3 (2000).

Abecasis, G. R., Chemy, S. S. & Cardon, L. R. The impact of genotyping error on
family-based analysis of quantitative traits. EurJ Hum Genet 9, 130-4 (2001).

Douglas, J. A., Boehnke, M. & Lange, K. A multipoint method for detecting
genotyping errors and mutations in sibling-pair linkage data. Am J Hum Genet 66,
1287-97 (2000).

Stringham, H. M. & Boehnke, M. Identifying marker typing incompatibilities in
linkage analysis. Am J Hum Genet 59, 946-50 (1996).

Sobel, E., Papp, J. C. & Lange, K. Detection and integration of genotyping errors
in statistical genetics. Am J Hum Genet 70, 496-508 (2002).

Douglas, J. A., Skol, A. D. & Boehnke, M. Probability of detection of genotyping

errors and mutations as inheritance inconsistencies in nuclear-family data. Am J
Hum Genet 70, 487-95 (2002).

122

   

11111111111111111111