Hf.

m. V

u:
an?

r ...
.z.
1,1...‘Cdzi. , , , . . . HM! .. .... . .. l,
.. .Fan . ‘ ‘ . . , , . , 1...: a 1
~ . . . . . , , x :r: 1 . 1p
. V . . nun. ‘ 1a.;
, . ‘ . . .11“
‘.:z a.«. . . . . . . V .
«”2“. 3‘. ‘ .
e
4

V Wag)- .r
x4..huv

.3

 

.Ltnv.)

5 m .1 1
s8... 3“:

 

I. ‘iiiw: ‘ .,
5.... 1118'. Tum“
l... .1». 34. I.
521...} .2

 

 

 

 

Warn}:
31:5. .0 .3.)
5151 x if:

 

h .539?!“

vvrhA Vin/A
‘..»“;:J
.;I.

 

. ....... .Ho. .uvflgflwawxzéflfi

 

 

 

This is to certify that the
dissertation entitled

IN UTERO PROGRAMMING OF CHILDHOOD ASTHMA

presented by

Alireza Sadeghnejad

has been accepted towards fulfillment
of the requirements for the

PhD. degree in Epidemiology

 

 

(MM Lfl

 

Major Professor’s Signature
( ,

Date

MSU is an Affirmative Action/Equal Opportunity Institution

 

LIBRARY
MiChiQan State
University

 

 

 

 

 

-.-.- -.- —..—.-.-t-.-.—o- -.— ¢- -

- -.-.- -..- -.-_.-

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 p:/ClRC/DateDue.indd-p.1

IN UTERO PROGRAMMING OF CHILDHOOD ASTHMA

By

Alireza Sadeghnejad

A DISSERTATION
Submitted to
Michigan State University
In partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
DEPARTMENT OF EPIDEMIOLOGY

2006

ABSTRACT
IN UTERO PROGRAMMING OF CHILDHOOD ASTHMA
By
Alireza Sadeghnejad
Asthma, the most common chronic illness of childhood, is a public

health problem and its prevalence has increased in the last decades. The
purpose of this dissertation was to investigate whether maternal exposures
during pregnancy can increase the risk of childhood asthma in offspring, a
process known as in utero programming. To address this concept, a
theoretical model with three components was proposed:

1- Gene-environment interactions during pregnancy affect a marker

(susceptibility marker) that is involved in asthma pathogenesis.

2- The susceptibility marker, detected at birth, has a long-lasting

effect on asthma.

3- The combination of genetic predisposition and exposures during

pregnancy affect asthma later in life.

lnforrnation on maternal smoking during pregnancy, the interleukin 13

gene (IL13), cord serum immunoglobulin E (CS-lgE, a susceptibility marker),
and asthma phenotypes were used to build the theoretical model for the in
utero programming process. The data used in the analyses is from the Isle of
Wight (United Kingdom) birth cohort, established in 1989, and comprised of
1,456 children who have been thoroughly characterized for asthma

phenotypes and other allergic diseases as well as allergic markers at birth, 1,

2, 4 and 10 years of age.

The results suggests first, a modifying effect of maternal smoking on the
association between the common variant of the IL13 gene and levels of cord
serum immunoglobulin E (CS-lgE); second, an association between CS-lgE
and childhood asthma; third, a combined effect of the common IL13 variant
and maternal smoking during pregnancy on early onset persistent wheeze
phenotype. The three pieces of this work suggest that in utero exposure to
tobacco smoke may alter the expression of the IL13 gene at birth that lasts as
late as age 10. The interaction of IL 13 polymorphism was not observed in
those who were exposed to tobacco smoke after birth unless they had
exposure during pregnancy. This suggests that the in utero environment has
a critical effect on the pathogenesis of asthma and that the prenatal period is

of utmost importance.

It is of public health importance to consider the prenatal period as a
critical time in the pathogenesis of diseases and to start education for
exposure prevention before a pregnancy happens. Future studies may
explore the molecular basis for the interaction between prenatal exposure to
tobacco smoke (and other pollutants) and genes. lf epigenetic changes play a
role in the pathogenesis of asthma, it may even be possible to reverse the

effect of pollutants through intervention therapies.

DEDICATION

This work is dedicated to my parents, Negin and Eilia.

iv

ACKNOWLEDGMENTS
The planning, designing, implementation and completion of this work is
the result of dedicated individuals. Sincere gratitude to my dissertation
advisor, Dr. Wilfried Karmaus, for his insight, supervision and moral

support in the process of completing this dissertation.

I would also like to thank the other members of my committee, Drs.
Suasn Ewart, Venu Gangur and Joseph Gardiner for their support during
this project.

Thanks to Mrs. Susan Davis for her meticulous editorial comments.

TABLE OF CONTENTS

LIST OF TABLES ........................................................................................... vii
LIST OF FIGURES ....................................................................................... viii
LIST OF ABBREVIATIONS. .......................................................................... ix
INTRODUCTION ............................................................................................ 1
METHODS .................................................................................................... 21
RESULTS ..................................................................................................... 38
DISCUSSION .............................................................................................. 107
APPENDIX ............................................................................................................ 1 15
REFRENCES .............................................................................................. 129

vi

LIST OF TABLES

TABLE 1.1 .................................................................................................... 20
TABLE 2.1 ................................................................................................... 25
TABLE 2.2. .................................................................................................. 27
TABLE 2.3. ................................................................................................... 37
TABLE 3.1.1 ................................................................................................. 53
TABLE 3.1.2 ................................................................................................. 55
TABLE 3.1.3 ................................................................................................. 56
TABLE 3.1.4 ................................................................................................. 57
TABLE 3.1.5 ................................................................................................. 58
TABLE 3.2.1 ................................................................................................. 77
TABLE 3.2.2 ................................................................................................. 78
TABLE 3.2.3 ................................................................................................. 79
TABLE 3.2.4. ............................................................................................... 80
TABLE 3.2.5. ............................................................................................... 81
TABLE 3.3.1 ............................................................................................... 100
TABLE 3.3.2 ............................................................................................... 101
TABLE 3.3.3 ............................................................................................... 102
TABLE 3.3.4 ............................................................................................... 103

vii

LIST OF FIGURES

FIGURE 1.1 .................................................................................................. 16
FIGURE 2.1. ................................................................................................ 22
FIGURE 2.2. ................................................................................................ 28
FIGURE 2.3 .................................................................................................. 34
FIGURE 3.1.1 ............................................................................................... 59
FIGURE 3.1.2 ............................................................................................... 60
FIGURE 3.2.1. ............................................................................................. 82
FIGURE 3.2.2. ............................................................................................. 83
FIGURE 3.3.1. ........................................................................................... 104
FIGURE 3.3.2. ........................................................................................... 105

viii

LIST OF ABBREVIATIONS

CI ........................................................................... CONFIDENCE INTERVAL
lgE .............................................................................. IMMUNOGLOBULIN E
lL .............................................................................................. INTERLEUKIN
DNA ................................................................... DEOXYRIBONUCLEIC ACID
OR .......................................................................................... ooos RATIOS
SAS ......................................................... STATISTICAL ANALYSIS SYSTEM
SNP .............................................. SINGLE NUCLEOTIDE POLYMORPHISM
SPT ................................................................................... SKIN PRICK TEST
,3 ............................................................................................... CHI-SQUARE

ix

INTRODUCTION

1.1. Overview

Asthma is a public health problem, the most common chronic illness of
childhood, with a prevalence that has increased in the last decades. The purpose
of this dissertation is to investigate whether maternal exposures during
pregnancy can increase the risk of childhood asthma in offspring, a process
known as in utero programming. To address this concept, a theoretical model is
proposed. The three components of this model are:

1- Gene-environment interactions during pregnancy affect a marker

(susceptibility marker) that is involved in asthma pathogenesis.

2- The susceptibility marker, detected at birth, has a long-lasting effect on

asthma.

3- The combination of genetic predisposition and exposures during

pregnancy affect asthma later in life.

Information on maternal smoking during pregnancy, the interteukin 13
gene (IL 13), cord serum immunoglobulin E (CS-lgE, a susceptibility marker), and
asthma phenotypes were used to build the theoretical model for the in utero
programming process. The data used in the analyses are from the Isle of Wight
(United Kingdom) birth cohort, established in 1989, and comprised of 1,456
children who have been thoroughly characterized for asthma phenotypes and
other allergic diseases as well as allergic markers at birth, 1, 2, 4 and 10 years of
age. The Institutional Review Board at Michigan State University approved this

work.

In this first chapter, I give the definition and pathophysiology of asthma; its
public health significance; and the effects of exposure to tobacco smoke, genetic
predisposition, and their interactions on asthma. Next, the importance of time
windows of exposures and the main concepts that try to explain the increase in
the prevalence of asthma are introduced. Then the theoretical model, and its
three components (see above) and their backgrounds are presented. In
‘Methods’ I describe the methods that are not discussed thoroughly in the
manuscripts that are available in the ‘Results’ section. More details on methods
can be found in the manuscripts. The ‘Results’ section is composed of three
manuscripts, one published and two submitted for publication, and provides
findings related to the following questions:

1- IS there an interaction between inteneukin 13 gene (IL13) and maternal

smoking during pregnancy on CS-lgE?

2- Is CS-lgE related to childhood asthma?

3- Is there a combined effect of IL 13 polymorphism and maternal smoking

during pregnancy on childhood asthma phenotypes?

In the last section, ‘Discussion and Conclusion’ the results are discussed,
conclusions are presented, practical issues regarding public health are
addressed, and future scientific steps toward the early life prevention of
Childhood asthma are offered. Appendices comprise a glossary and logs from the

statistical programs SAS and PHASE used in analyses.

1.2. Definition and pathophysiology of asthma

Asthma is a chronic inflammatory disorder of the lungs that is Characterized
by a reversible obstruction of the airways. (1, 2) Allergy, an immune response to
allergens, is characterized by an increased level of lgE, and can be one of the
elements of asthma (typical allergen sources include house dust mite, grass
pollens and animal dander). (3) Classically two different pathologic entities have
been described, allergic and non-allergic asthma, although the differences are
not yet clearly understood. (4, 5) From pathophysiologic aspect, asthma is
characterized by a variable ainivay obstruction due to ainivay inflammation and
increased bronchial responsiveness to different external and internal stimuli. The
inflammation results in mucosal edema, cellular infiltrate, and glandular
hyperplasia, whereas bronchial hyper-responsiveness (BHR) causes excessive
tone and contraction of the bronchial smooth muscles. Both of these processes
lead to an intermittent narrowing of the airways. This narrowing produces
"wheeze”, which is considered to be the clinical hallmark of asthma. The immune
response that produces inflammation in asthma is driven by a predominance of
cytokines: interleukin (lL)—4, lL-5, and lL-13. It has been suggested that immune
dysregulation causes inflammation with resultant airway structural damage and
repair (remodeling). (6) The structural changes observed include thickening of
the reticular basement membrane, an increase in the underlying collagen, and
hypertrophy of the airway smooth muscle. Abnormally thickened airways may be
the mechanism underlying both BHR and fixed airway obstruction. (7) These

changes begin in childhood and underscore the importance of early detection

and the appropriate intervention for persistent wheeze and asthma. (8) Thus, it is
important to study the natural history of Childhood asthma and identify risk factors
associated with its development.

One challenge in studying asthma is the presence of enormous variability in
the features of asthma (including symptoms and lung function) both between and
within patients over time. For genetic studies, the phenotype associated with
asthma can be defined by subjective measures (e.g., symptoms), objective
measures (e.g., airway hyper-responsiveness or serum lgE level), or both.
Because of the complex Clinical presentation of asthma, the genetic basis of the
disease is often studied through intermediate phenotypes that can be measured
objectively, such as the presence of sensitization or airway hyper-
responsiveness. However, the latter phenotypes are not specific to asthma, and
they may lead to a false positive diagnosis of asthma. (9, 10) This lack of a clear
definition of the asthma phenotype is an obstacle for studies of the genetic basis
of asthma. (11-13)

Epidemiological definitions of asthma are based on symptoms, tests for
ainrvay hyper-responsiveness, and an evaluation of allergy. Questionnaire
information has formed the basis of many epidemiological studies and has been
used to define asthma in terms of symptoms alone. Asthma and some wheeze
phenotypes, like “persistent wheeze’, have been used interchangeably.
Commonly used definitions of asthma are based on “wheeze ever" and “wheeze
in the past 12 months”. (2) The Tucson Children’s Respiratory Cohort Study has

identified childhood wheezing phenotypes with differing Characteristics and

prognoses, such as early transient wheeze, non-atopic wheeze, lgE-mediated
wheeze, and asthma. (14) The Isle of Wight birth cohort study has extended
these observations, describing characteristics of childhood wheezing phenotypes
and identifying risk factors associated with their development. (15, 16)

1.3. Public health significance of asthma

Asthma is a serious global public health problem and it is the most
common chronic illness of childhood. The prevalence of asthma symptoms in
children is different in different populations and can be as high as 30 percent.
The highest prevalences are found in Australia, New Zealand, and England.
Childhood asthma is more common in boys than in gins. Asthma is a disease
that places a huge burden on society. It is estimated that asthma affects more
than 15 million persons in the US, leading to more than 500,000 hospitalizations
and over 5,000 deaths annually. (17) The annual direct cost of asthma was
recently estimated to be 8 billion dollars. (18)

There is evidence that the prevalence of asthma and other allergic
diseases has been rising at a dramatic rate in the last three decades. It is also
becoming clear that this rise is largely restricted to developed countries, and is
widely believed to be linked to a Western lifestyle. In the United States, the
prevalence of asthma has increased by nearly 75% in the past 2 decades. (19)
This recent increase has been attributed to environmental factors and their
interactions with genes. (5, 20-24) Exposure to tobacco smoke is one of the
major risk factors for asthma that has been reported in a large number of studies

and it is a focus of this dissertation.

1.4. Exposure to tobacco smoke

Among environmental exposures, the association of tobacco smoke and
asthma has been investigated to some extent, and there is evidence that both
maternal smoking during pregnancy and passive smoke exposure after
pregnancy affect asthma later in life. (25-30) Using data from a large cohort study
(n~60,000), Jaakkola et al. demonstrated that maternal smoking in pregnancy
increases the childhood risk of asthma. (22)

Using the information from more than 20,000 children (6 to 12 years of
age) from nine countries in Europe and North America, Moshammer et al.
reported that smoking during pregnancy decreases lung function parameters.
(31) The effects of past and current passive exposure to tobacco smoke were
smaller than the effect of smoking during pregnancy. This study strongly
suggests that maternal smoking during pregnancy subsequently affects the lung
function of Children.

Despite the evidence for the effect of prenatal exposure to tobacco smoke
on asthma phenotypes, there is very limited knowledge about the mechanism of
this effect. (22) This dissertation investigates whether the effect of prenatal
exposure to tobacco smoke on asthma phenotypes later in life iS related to a
genetic predisposition (IL13 gene polymorphism).

1.5. Genetic predisposition

There is evidence that asthma has genetic components. (2) A number of

studies have shown an increased prevalence of asthma phenotypes among the

offspring of subjects with asthma compared to the offspring of subjects without

asthma. (11, 12) Family studies have convincingly indicated that atopy (as
measured by allergen Skin tests, total lgE, and/or specific lgE), airway hyper-
responsiveness, and asthma as diagnosed by questionnaire are at least partly
under genetic control. (11, 32) Numerous studies of twins have demonstrated
that concordance rates for asthma phenotypes are all substantially higher for
monozygotic than for dizygotic twins, suggesting a strong genetic contribution. ln
population-based studies of twins, the estimated effect of genetic factors is about
35 to 70 percent, depending on the population and the design of the study. (11,
12) However, despite extensive effort and advances in molecular biology and
genetics, no gene for asthma has been identified with any certainty. (11, 12) The
results of several studies provide an indication that multiple genes may be
involved in the pathogenesis of asthma, and chromosomal regions likely to
harbor asthma susceptibility genes have been identified. (12, 33, 34) Several
genes on chromosome 5q may be important in the development or progression
of inflammation associated with asthma and atopy, including genes encoding the
cytokines interleukin (lL)-3, lL-4, lL-5, IL-9, lL-12, and lL-13. (4,5) In particular, IL-
13 plays an essential role in the allergic immune response, both by causing the
differentiation of T-helper cells toward an allergic response and by inducing lgE
production by B cells. (35, 36) lL-13 also affects both ciliated and secretory cell
differentiation, leading to ainNay damage and obstruction. (37) The other effect of
lL-13 that is very important in the pathogenesis of asthma is increased mucus
production in the respiratory tract. (38) Therefore, IL13 is a strong candidate

gene for having a role in asthma. In this work information on IL13 gene

polymorphism (see ‘methods’ section) is used to investigate its combined effect
(interaction) with maternal smoking during pregnancy on the outcomes of
interest.

1.6. Gene-environment interaction

It has been suggested that common diseases, such as diabetes, cancer,
and asthma, are the result of a complex interplay of genetic and environmental
factors. The concept of gene-environment interaction implies that a combination
of the two factors (a gene and an environmental exposure) has a different effect
than their independent impacts. Statistically, a gene-environment interaction can
be tested in several ways. One way is to stratify the sample by one factor and
look at the effect of the other. Another way is to explore the statistical
significance of a gene-environment interaction by including terms for the gene,
the environmental factor, and their interaction in a model. (39)

The fact that no single gene has been identified for asthma, and that the
recent increase in the prevalence of asthma cannot be explained by genes, has
prompted researchers to investigate gene—environment interaction effects for
asthma. In consideration of the objectives of this dissertation, I will address the
following issues:

1- Evidence for the combined effect of genetic predisposition and

exposure to tobacco smoke on asthma

2- The epigenetic explanation as one of the mechanisms for gene-

environment interactions

3- The importance of a time window for a gene-environment

interaction

4- The prenatal programming concept
1.6.1 . Evidence for the combined effect of genetic predisposition and
exposure to tobacco smoke on asthma

For asthma, several studies have Shown a combined effect of genetic
predisposition and exposure to tobacco smoke. (30, 40, 41) Wang et al.
observed that the association of [32 adrenergic receptor genotypes and asthma is
affected by exposure to tobacco smoke. Kabesch et al. reported that the effect
of passive smoking on asthma was larger when children had glutathione S
transferase deficiency. A genome-wide analysis has suggested more links
between genetic markers and asthma when subjects were exposed to passive
tobacco exposure. (30)
1.6.2. Epigenetic phenomena as one of the mechanisms for gene-
envlronment interactions

One mechanism used to explain the effect of gene-environment
interactions on asthma is known as epigenetics. (5) All cells of an individual have
a unique composition of DNA and hence, genes. The process of a gene
becoming active (turned on) is known as gene expression. The structural and
functional heterogeneity of cells is the result of differential patterns of gene
expression. A mutation is a change in the DNA sequence. A mutation of a gene
can potentially alter gene expression, but mutation is a rare event. Usually

alteration in gene expression happens without mutations in the DNA sequence

10

and rest on the potential ability of genetic information to change through
developmental events and/or environmental exposures. These changes happen
via the post synthetic modifications of either the structure of the double strand
helix (by methylation) or of histones (by acetylation, methylation, or
phosphorylation) (42). Based on animal studies, it is evident that the methylation
status of DNA and the chromatin conformation are affected by environmental
exposures (42, 43). These stable alterations in genetic information that are
heritable in the short term and do not involve a sequence variation (mutation) in
the DNA itself, are called epigenetic changes. If a modification of the genetic
information occurs during early fetal life, it may be retained through cell division
(mitosis) (5, 42, 43) and over generations. (44, 45) This possible life long impact
highlights the importance of a time-window for environmental exposures in the
pathogenesis of asthma.
1 .6.3. The importance of a time window for gene-environment interaction
Regarding the effect of exposures on asthma and its recent increase in
prevalence, there are two concepts: First, the hygiene concept, and second, the
prenatal priming concept. The first concept emphasizes the time window after
birth as a critical period of exposure and considers that a decline in the exposure
to microbial antigens (hygiene) is responsible for the recent increase in allergic
disorders. Supporters of this concept suggest that improved hygiene in
industrialized societies and the use of vaccines and antibiotics have reduced the
incidence of infections. The assumption is that cytokines involved in immune

responses against infection can suppress immune responses that are involved in

11

allergic reactions. Therefore, a reduced exposure to infectious agents can result
in an augmented allergic response and, hence, allergy. (46, 47) The studies that
suggest this concept are epidemiological investigations that Show the following:
an inverse relationship between family size and the risk of developing asthma;
early placement in day care settings, and presumed exposure to infectious
agents, appears to protect against the development of asthma; and, observations
demonstrating that exposure to farm animals and raw milk early in life reduces
the likelihood of developing asthma, allergic rhinitis and atopic sensitization. (48-
50). The studies of the hygiene concept are contradictory and also they fail to
provide a causal relationship between decreased exposure to infectious agents
and allergy. (51, 52) The second concept (prenatal priming, the focus of this
dissertation) considers fetal life as the critical period of exposure and suggests
that exposure to environmental factors during pregnancy are important for the
rise in the prevalence of asthma and allergy. (53-55) A more detailed discussion
of prenatal programming follows.
1.6.4. The prenatal programming concept

During organogenesis in fetal life, tissues face ‘critical’ periods of
development. ‘Programming’ describes the process whereby a factor at a critical
period of development has lasting or lifelong effects. (56) AS an example, it has
been shown that offspring of women who were pregnant when they experienced
the Dutch Hunger Winter of 1944-1945 had more insulin resistance and Type II
diabetes than those born before or conceived after the famine. (57) Although the

concept of “in utero programming" has been investigated for many chronic

12

disorders, the previous studies mainly focused on fetal undemutrition, without
considering a Specific factor or exposure. For asthma and allergic diseases, the
role of the uterine environment in programming the immune responses of the
infant has not been widely studied and allergen exposure after birth has been
considered as the major factor in whether a baby at high genetic risk of atopy
develops disease. (58)

It is generally accepted that dramatic Changes in maternal immune
responses occur during pregnancy. (59) Fetal markers of an immune response
are detected as early as 10-22 weeks gestation. (60) T-cells have essential roles
in the ways that the immune system responds to infectious agents and antigens.
Among the different types of T-cells, Th1 cells react to intracellular infectious
agents and foreign cells; Th2 cells are involved in antibody production and
allergic reactions. The type of T-cell reaction that is related to allergic outcomes
(T h2) is also enhanced in the feto-matemal unit during gestation. (61) The
reason for Th2 enhancement during pregnancy is to diminish reactions of
maternal and fetal immune systems against each other that may terminate
pregnancy. (62, 63) Although the above-mentioned dichotomy for T-cells has
been challenged recently, immune reactions in allergy and in a normal pregnancy
are considered to be similar. (64) The latter suggests that the prenatal period is a
critical time window for the development of allergic diseases. There is also
evidence that maternal immune status and exposures during pregnancy may
change the susceptibility of the mother’s fetus to childhood atopic disorders. (5)

Furthermore, antigens that can be transferred to the fetus during pregnancy may

13

play a role in the future development of atopy. There is, in fact, evidence that
exposure to antigens during pregnancy can affect the fetal immune response.
(65, 66) For example, Van Duren-Schmidt et al. showed that proliferative
responses of umbilical cord blood mononuclear cells (CBMC) to birch pollen and
timothy grass pollen were higher when the amount of exposure to these pollens
was higher during fetal life. (67) CBMC responses to specific allergens are
indicators that in utero exposure to factors that have entered the maternal
circulation during pregnancy results in the development of early infant responses
(58, 66), which may be predictive of later allergic disease. Additionally, as
discussed previously, gene-environment interactions through epigenetic
mechanisms can have a profound effect during fetal life compared to the
postnatal period.

CS-lgE has been suggested as a possible marker of exposure to
environmental factors during pregnancy that may be used to predict allergy later
in life. Several groups investigating the predictive capacity of elevated CS-lgE for
the development of allergic disease have demonstrated that lgE specific for
certain allergens can be detected at birth. (68, 69) Therefore, the in utero
environment influences fetal responses in a way that may have major
implications later in life.

Despite the importance of the prenatal period as the critical time window
for the effects of environmental exposures, there is very limited information for
the effect of interactions between genes and exposures during pregnancy on

Childhood asthma. Based on the prenatal programming concept, I conceptualized

14

a theoretical model. Using the interleukin 13 gene (IL13) and maternal smoking
during pregnancy as examples for the theoretical model, this dissertation
investigates whether there is an in utero programming process for the
development of asthma in offspring.
1.7. Theoretical model

For an in utero programming process for childhood asthma to exist, I
deem that three conditions need to be met: First, presence of evidence that a
biologic mechanism can change the susceptibility to asthma. Cord serum
immunoglobulin E was used as the marker of susceptibility for asthma (CS-lgE of
course, would not be affected by after birth exposures; and could potentially be
an indicator of in utero influences). The biologic mechanism for programming
was hypothesized to be a gene-environment interaction between the IL13 and
maternal smoking during pregnancy; Second, existence of an association
between the marker of susceptibility at birth, CS-lgE, and asthma phenotypes
later in life. This association suggests that some stable risk of asthma can be
attributed to fetal life events; Third, evidence is required that the biologic
mechanism during fetal life (the IL13 genotypes - maternal smoking during
pregnancy interaction) can affect asthma phenotypes later in life. The
requirement of a similar gene-environment interaction for both CS-lgE and
asthma provides further support for the existence of a biological process rather
than mere statistical associations. The three associations for the overarching
theme of an in utero programming process are Shown in figure 1 and discussed

in the background section (below).

15

Prenatal < Bin" > Postnatal

pefiod pefiod

IL13 x maternal
smoking

   

03495 Asthma
O @ phenotypes

  

 

Figure 1.1. The three associations for the overarching theme of an in utero
programming process.

Interleukin 13 gene (IL13), maternal smoking, cord serum immunoglobulin
E (CS-lgE) and asthma phenotypes are examples used in this
dissertation. 1- The interaction between IL13 and maternal smoking
affects CS-lgE (a susceptibility factor for asthma) 2- CS-lgE is related to
asthma later in life 3- The interaction between IL13 and maternal smoking
is present for asthma phenotypes.

16

1.8. Background
1.8.1. The effect of maternal smoking during pregnancy and IL13 gene
polymorphism on cord serum immunoglobulin E.

Allergic reactions have long been recognized to be lgE-mediated. (70, 71)
Evidence suggests that factors early in life have major effects on the later
development of allergy, and cord serum lgE (CS-lgE) has been suggested to be
a biomarker for allergic diseases later in life. Several reports have shown a
relationship between IL13 and serum lgE (72, 73), but to the best of my
knowledge, there has been no investigation of the effect of IL13 polymorphisms
or their interaction with environmental exposures on CS-lgE. In fact there is very
limited information on the relationship between genes and CS-lgE. (74) Chang et
al. reported a gender-limited association of cytotoxic T-lymphocyte antigen-4
(CTLA-4) polymorphism with cord blood lgE levels. The results of studies of the
effect of maternal smoking on CS-lgE have been inconsistent. (75-81). If we
assume that the genetic background is different across populations, and if the
effect of cigarette smoking on CS-lgE is dependent on the genetic background,

then some of this inconsistency can be attributable to gene-smoking interactions.

17

1.8.2. Cord serum lgE is related to childhood asthma.

Orgel et al., 1975, was first to report a correlation between serum lgE
during the first year of life and atopic diseases in the first 2 years of life. (82) As
mentioned above, CS-lgE has been suggested to be a biomarker for atopic
diseases later in life. (78, 82-88) However, there are studies that disagree with
the use of CS-lgE as a biomarker for asthma and allergy. (68, 83, 88-93) A
limitation of these studies is that only early childhood manifestations (up to 5
years) were investigated. (78, 82, 90, 94) Except Nambu et al., who investigated
CS-lgE Specific to a particular allergen, all of the previous studies used total CS-
lgE. (83) Comparisons among previous studies are complicated by their use of
different allergic outcomes. For this dissertation the association between CS-lgE
levels and asthma up to age 10 was investigated.

1.8.3. The combined effect of maternal smoking during pregnancy and IL13
gene polymorphism on childhood wheeze phenotypes.

It has been suggested that the increased prevalence of asthma-related
phenotypes over the last three decades is due to exposure to environmental
factors (95). Among such exposures, cigarette smoking is regarded as an
important risk factor for asthma (22). In particular, an effect of maternal smoking
during pregnancy on the development of asthma in offspring has been proposed
(96-102). A number of studies in children have linked parental smoking with
markers of allergy such as serum lgE levels (103, 104). Interleukin 13 (IL-13) is
an important cytokine involved in the lgE pathway and pathogenesis of asthma

(35, 105, 106). Noakes et al. noted that exposure to tobacco smoke during

18

pregnancy was associated with a significantly higher neonatal lL-13 response to
both house dust mite and ovalbumin, indicating that a prenatal exposure to
tobacco may have immunological effects (105). These observations suggest that
part of the effect of exposure to tobacco on asthma might be through cytokines.

Researchers have speculated that individuals may vary in genetic
susceptibility to cigarette smoke exposure (104, 107). A few studies have
suggested a combined effect (interaction) of genes and exposure to tobacco for
asthma (30, 40, 41). In Spite of the importance of lL-13 in asthma (35, 105, 106),
some studies have failed to Show an association between the interleukin 13 gene
(IL 13) polymorphism and asthma phenotypes (108, 109) and there is no
information as to whether the adverse effect of maternal smoking on asthma may
result from an interaction between exposure to tobacco smoke and cytokine
regulating genes. To the best of my knowledge, no study has investigated
whether there is an interaction between tobacco smoke exposure and IL13
polymorphism on wheezing. Because of the high prevalence (25-30 percent) of
smoking among women of reproductive age (110) there is a need to investigate
its possible role in the development of asthma and related phenotypes. Here, I
investigate the presence of an interaction between exposure to tobacco smoke
and IL13 gene polymorphism on longitudinal assessments of wheeze

phenotypes during the first decade of life in the Isle of Wight birth cohort.

19

Table 1.1. Studies which investigated the association between cord serum
immunoglobulin E (CS-lgE) and allergy later in life

 

 

Author, year Cut off point for Sample Age (years) Association with
CS-IgE (kU/L) size at follow up Asthma phenotypes
Nambu, 2003 0.07 10 3 No
Kaan, 2000 0.5 384 1 No
Bergmann, 1997 0.35 1314 2 No
Kjellman, 1994 0.9, 1.3 1654 11 Yes
Eiriksson, 1994 0.2, 0.7 180 1.5- 2 No
Ruiz, 1991 0.7 92 1 No
Magnusson, 1988 1.20 190 1.5 No
Strimas, 1988 0.5 83 1 No
Croner, 1982 1.3 1701 1.5 No

 

20

METHODS

21

In this chapter, I describe the Isle of Wight Birth Cohort Study and
methods used for statistical and genetic analyses. Methods, specific to each
research question are available in the three manuscripts in the ‘Results’ section.
2.1. The Isle of Wight Birth Cohort Study

The Isle of Wight Birth Cohort
Study represents an unselected whole 0;
population birth cohort based on the Isle E:

/ W)
A
of Wight, UK. The Isle of Wight is an M

‘71
island (21 x 37 km) off the southern g“? Egg;
r 0
coast of England with a resident J
cm?” K
population of 130,000. The majority of & "\\
. . . United \
people live In small towns and Villages. Kingdor’fi"\“\
The population density is similar to that S N; >
of the rest of England (340 per sq km). 3 EL fl
. . . ”5ft:- ,9 London
There Is no major manufactunng or W\__,,~~.// £~a§ :
‘17ng Somiw‘)
large industries on the island. In 2001 ./ mfg?
f Isle of Wight

only 2.3% of residents were employed in (g,

Figure 2.1. Isle of Wight is located

angculture trade. Thus, the Isle of Wight in the south of United Kingdom

is primarily a semi-rural but not a
farming community. The population is mainly Caucasian (99%) and it is neither
geographically nor genetically isolated.

The Isle of Wight birth cohort study was initiated by Drs. David Hide and

Hasan Arshad at the David Hide Asthma and Allergy Research Centre on the Isle

22

of Wight. The intention was to prospectively study a whole population cohort for
the development of asthma and allergic diseases and identify genetic and
environmental risk factors relevant to these conditions. Enrollment took place at
birth. Of the 1,536 children born on the Isle of Wight between January 1, 1989
and February 28, 1990, informed consent was obtained from the parents of 1,456
children. These children have since been seen at the ages of 1, 2, 4 and 10
years. Detailed information on family history and environmental exposures have
been collected at birth and updated at each follow-up.

At each follow-up, detailed questionnaires were completed with the
parents for each child regarding asthma and allergy prevalence. Information on
asthma morbidity and medication requirement was also obtained. Skin prick
testing to 14 common food and aero-allergens was performed at 4 and 10 years.
Further information was collected at 10 years using standardized lntemational
Study of Asthma and Allergy in Childhood (ISAAC) questionnaires. (111)
Spirometry and bronchial provocation tests were performed, blood was collected
for genetic studies, and serum was used for lgE measurements and stored for
future analysis.

The Research Ethics Committees of the David Hide Asthma and Allergy
Research Center and Michigan State University approved the study of Children
born and enrolled (n=1,456) between January 1, 1989, and February 28, 1990,
on the Isle of Wight. Informed written consent was obtained from parents.
Children were followed up at the ages of 1 (n = 1,167; 80.2%), 2 (n = 1,174; 80.6

percent), 4 (n = 1,218; 83.7%) and 10 years (n = 1,373; 94.3%).

23

2.2. Statistical analysis
SAS/STAT® program versions 8.2 and 9.1 were used to perform

statistical analysis. (112) Statistical significance was defined as a p value of s
0.05. The analyses include Chi-square (38) tests, likelihood ratios, logistic
regression, and assessment of statistical interactions.

2.2.1. x2 distribution and tests

The Chi-square (12) test is a statistical procedure to assess association

between two categorical variables. When the variables are arrayed in r-rows and

c-columns, the test of the null hypothesis of no association has degrees of
freedom (r-1)(c-1). The 12 test can be used when no more than 20% of the

expected counts are less than 5 and all individual expected counts are 1 or
greater. For a 2X2 table, all four expected counts should be 5 or greater. (113)

Formula for getting the expected frequency of a particular cell

. 1St column, 1St row = [(row 1 total) x (column 1 total)]/ grand total
. 1St column, 2nd row = [(row 1 total) x (column 2 total)]/ grand total

0 In general, expected frequency of ith column, jth row=[(row i total) x
(column j tota|)]/ grand total

In this dissertation 38 tests were used to compare the sample used in each

analysis with the reference sample. Consider that Sample 1 is the original group

and Sample 2 is the one used in the analysis and number of missing is M (table
2.1). Let us review the previous exercise that about 12 test for equality of

proportion

24

Table 2.1. A scenario to investigate whether a missing
process is related to the different values of a variable

Variable of interest Value 1 Value 2 TOTAL

 

Sample 2 a b a + b
Sample 1- Sample 2 M1 M2 M
TOTAL (Sample 1) A B A + B

 

The question is whether a significant difference between the proportions of
missing by the value of the variable exists. We test:

Ho: p1=p2 vs H1: pt¢p2

where, p1: proportion of missing for Value 1 and

p2: proportion of missing for Value 2

2.2.2. Likelihood ratios

The likelihood ratio is defined as: sensitivity / (1-specificity). These ratios
are used to compare cut-off points of a marker, for example CS-lgE, and choose
the cut-off point that provides the optimal sensitivity and specificity for the
outcome of interest, for example asthma. This method is used in the second
manuscript in the ‘Results’ section.
2.2.3. Logistic regression

Logistic regression is primarily used to assess the relationship to the
binary outcome of a primary exposure, after adjusting for the effects of other
variables (covariates) and their interaction with the exposure. When these

variables are not known to be in the causal pathway of association between

25

exposure and the outcome they are known as confounders. Confounders that
interact with the primary exposure are called effect-modifiers.

Let Y denote the dichotomous outcome of interest with Y =1 labeling the
presence of the condition under study, and Y=0 labeling the absence of the

condition. Covariates, including the primary exposure and interactions are
denoted by X1, X2. . X... Then the logistic model for P [Y=1 | X1, X2. . .Xk] is written
simply as P03) where _)_( is abbreviation for X1, X2. . X... The model formula is

given by

-(l30 +28i Xi)

 

P(_)$) = 1/ (1 + e )or logit (P()_()) = log ( ) = Bo +28, X, in

1 - p
which Bo and Bi are representing unknown parameters to be estimated based on
the observed variables X1 through Xk and the outcome variable Y. The
coefficients 81 through 8k can be interpreted as adjusted log-odds ratios. For
example, if X1 is the primary exposure, coded X1 = 1 for “exposed” and X1 =0 for
“not exposed”, then assuming the variables X2 through Xk do not include X1, the

odds ratio is given by exp(81).

2.2.4. Statistical interaction
2.2.4.1 . The concept of an interaction.

An interaction evaluates whether a change in one variable affects the
association between two other variables. In other words, it reveals whether or not
the effect of exposure on the outcome changes by different levels of a third

variable. For instance, consider two variants for IL 13 gene, major and minor; and

26

status of maternal smoking during pregnancy. When children had the minor IL13
variant, they had about 25% chance of having early onset persistent wheeze,
regardless of maternal smoking status. But when children had the major IL13
variant, maternal smoking increased their Chance of having early onset persistent
wheeze by 52% (41 .7+ 27.4, table 2.1). Graphically, interaction is present (Figure
2.2) when the slope for the effect of one factor (maternal smoking) changes by

different levels of the other factor (major and minor variants of IL13 gene).

Table 2.2. proportion of children who had early onset persistent
wheeze by different condition of interleukin 13 (IL13) genotype and
maternal smoking during pregnancy, Isle of Wight birth cohort study

Maternal smoked

 

 

no yes
minor 30/114(26.3) 10/40 (25.0)
IL 13 genotype
major 28/102 (27.4) 15/36 (41 .7)

 

27

 

 

 

-0- Major IL 13 genotype -0- Minor IL 13 genotype

 

50

 

\

30
5:0
'9.
:2
ow
::
.n
:0
K

 

20

10

 

 

Early onset persistent wheeze %

 

o , 1
Mother did not smoke during Mother smoked during pregnancy
Pregnancy

Figure 2.2. Graphical demonstration of a statistical interaction. The slope for the
effect of maternal smoking Changes by different variants (major and minor) of
interleukin 13 gene (IL13).

2.2.4.2. Evaluation of an interaction (additive) by stratification.

One way to test whether the dichotomous variable X1 (value ‘0’ for the
reference group and value ‘1’ for the comparison group) influences the
association between the dichotomous variable X2 (value ‘0’ for the reference
group and value ‘1’ for the comparison group) and the outcome of interest, Y, is
to stratify the sample by values of X1 and evaluate the association between the
other two variables, X2 and Y, in a logistic regression model. If the odds ratio of
X2 for Y is different when X1=0 versus when X1=1, it means that X1 may modify
the relationship of X2 and Y. This kind of interaction is known as additive

interaction.

It is also possible to test for the presence of a multiplicative interaction as follow.

28

2.2.4.3. Evaluation of an interaction (multiplicative) effect using logistic

regression.

To test the hypothesis that the dichotomous variable X1 (value ‘0’ for the
reference group and value ‘1’ for the comparison group) influences the effect of
the dichotomous variable X2 (value ‘0’ for the reference group and value ‘1’ for

the comparison group). and vice-versa, the model should include an interaction

term that is the product of these two variables. If the effect of one variable is
altered by the other, the coefficient on this interaction term (X1X2) will be

statistically significant.

In a logistic model, one can present the interaction between X1 and X2 as:
I09it (13(5)) = I30 + BIX1+ 52X2+ BSXIX2
To test whether a multiplicative interaction between X1 and X2 is present,

one can perform a likelihood ratio test.
2.2.4.4. Likelihood ratio test

First, the model with the interaction term is run and its likelihood is
recorded: Model 1: logit (Y) = I30 + B1X1+ B2X2+ 83X1X2 -) Likelihood=a

Second, the model without the interaction term is run and its likelihood is
recorded: Model 2: logit (Y) = 8’0 + B’1X1+ B’2X2 -) Likelihood=b

Third, as the difference between —2log (likelihood) of two nested (Model 1
and Model 2) has a x2 distribution under the null hypothesis that 83=0, one can
test the significance of the above-mentioned interaction by calculating p value for

the xz-test, in this case with one degree of freedom:

29

22 =-2log (a) - (-2log (b)) = -2|og (a/b)
It is also possible to determine the effect of an interaction:
For X1, the Change brought in logit (P()_()) by a Change in X1 from ‘0’ to ‘1’
while X2 = 1 is [31 + 82 as follow:
l09it (P11) = I30 + 131(1)“t 132(1)+ 53(1) (1) = I30 + 31* 52+ I33
logit (Poi) = I30 + BIIO)+ [32(1)+ (33(0) (1) = I30 + I31
logit (P11) - logit (P01) = 51 + I32
For X1, the change brought in logit (P(X» by a change in X1 from ‘0’ to ‘1’
while X2 = 0 is 81 as follow:
logit (Pio) = Bo + l31(1)+ l32(0)+ 33(1) (0) = Bo + I31
logit (Poo) = I30 + BI(O)+ 82(0)+ (33(0) (0) = Bo
logit (Pio) - logit (Poo) = 31
As presented, the effect of value 1 of X1 on outcome is different when X2 =
0 from when X2 = 1.

It iS also possible to calculate the combined effect for X1 and X2 as follow:
|091t(Pi1)=-' I30 + 51(1)+ [32(1)+ 53(1) (1) = I30 + 51* 132+ I33
logit (Poo) = I30 + BI(O)+ 82(0)+ 63(0) (0) = Bo
'09it (P11) - logit (Poo) = I31 1' 52 + 53

2.3. Genetic analysis

In this part three basic concepts are addressed:

1. Single nucleotide polymorphisms (SNPS)

30

2. Hardy-Weinberg equilibrium
3. Linkage Disequilibrium (LD) and tests for LD
2.3.1. Single nucleotide polymorphisms (SNPS)

Although more than 99% of human DNA sequences are the same across the
population, variations in DNA sequence can have a major impact on human’s
genetic structure and function. Therefore, markers of variations in DNA are of
great value for biomedical research. Single nucleotide polymorphisms (SNPS)
are DNA sequence variations that occur when a single nucleotide (A, T, C, or G)
in the genome sequence is altered. For example a SNP might change the DNA
sequence A_A_GGCTAA to AIGGCTAA. For a variation to be considered a SNP, it
must occur in at least 1% of the population. Usually a SNP needs a minor allele
frequency (MAF) of at least 10% to be useful in genetic studies. SNPS, which
could explain about 90% of all human genetic variation, occur every 100 to 300
bases along the 3-billion-base human genome. Two of every three SNPS involve
the replacement of cytosine (C) with thymine (T). SNPS can occur in both coding
(gene) and non-coding regions of the genome. Many SNPS have no obvious
effect (for example, they do not change an amino acid in the protein made by a
gene), but they still may change gene function (for example its expression). (114,
115)

Compared to other genetic markers, SNPS are evolutionarily stable, not
Changing much from generation to generation, so they are easier to follow in
population studies. Evolution is simply a Change in frequencies of alleles in the

gene pool of a population. For instance, let us assume that there is a trait that iS

31

determined by the inheritance of a gene with two alleles, B and b. If the parent
generation has 92% B and 8% b and their offspring collectively have 90% B and
10% b, evolution has occurred between the generations. The entire population's
gene pool has evolved in the direction of a higher frequency of the b allele—it was
not just those individuals who inherited the b allele who evolved. This definition of
evolution was developed largely as a result of independent work in the early 20th
century by Godfrey Hardy, an English mathematician, and Wilhelm Weinberg, a
German physician. (114, 115)
2.3.2. Hardy-Weinberg equilibrium

Hardy and Weinberg discovered an equation that can be used to
demonstrate the probable genotype frequencies in a population and to track their
changes from one generation to another. This is known as the Hardy-Weinberg
equilibrium equation: p2+2pq+q2=1

In this equation, p is defined as the frequency of the common (major)
allele and q as the frequency of the minor allele for a pair of alleles (A and a). In
other words, p equals all of the alleles in individuals who are major allele
homozygous (AA) and half of the alleles in people who are heterozygous (Aa) in

a population. In mathematical terms, this is
p = AA + 1/2Aa

Likewise, q equals all of the alleles in individuals who are homozygous for
minor allele (aa) and the other half of the alleles in people who are heterozygous

(Aa).

32

q = aa + 1/2Aa
Because there are only two alleles in this case, the frequency of one plus

the frequency of the other must equal 100%, which is to say:
p+q=1so.(p+Q)(p+q)=(p+q)2=1
Therefore, the chance of all possible combinations (p2 + 2pq + qz) of

alleles occurring randomly is 1. If the frequency of two alleles of one SNP do not
violate this equation (insignificant p value), the SNP is considered to be in Hardy-
Weinberg equilibrium. In other words the distribution of its alleles is not different
from the whole population. (114, 115)
2.3.3. Haplotype blocks

If SNPS are close together in the genome, the alleles of SNPS will tend to
be inherited together during meioses. The combinations of alleles of SNPS that
tend to stay together during meioses make haplotypes. The reason is that there
is less chance that recombination will separate the alleles of SNPS that are close
together. Regions corresponding to blocks have a few common haplotypes that
can explain most of the variability of a gene. As an example, for 2 SNPS, there

are four possible haplotypes as demonstrated in Figure 2.3.

33

   

Hflzkéh #1 b
a .3...—
a b

There are 4 possible haplotypes

\ B b

A pAB PAb
a p38 pab

 

 

Figure 2.3.For two SNPS there are four possible haplotypes (AB, Ab, a8 and
ab) with their corresponding probabilities (lower right).

When the genetic Information of parents is not available (unknown phase)
there are several algorithms and programs to estimate haplotypes and their
probabilities based on genotype data. One of the determinants to estimate
haplotypes is ‘linkage disequilibrium’ as follow.

2.3.4. Linkage Disequilibrium (LD) and tests for LD

LD is defined as the non-random association of alleles at two adjacent
SNPS. A pair of SNPS is considered to be in complete LD when their alleles are
fully associated. For example, consider two SNPS, X and Y with C-)T and A->G
variation, respectively. These SNPS are in complete LD when C at X appears
always with A at Y and T at X appears always with G at Y. Two tests, R2 and D’,
are frequently used to assess pair wise LD. For low allele frequencies R2 has

more reliable sample properties than D’. (116)

34

Suppose two bi-allelic loci, alleles A,a at locus 1, alleles B,b at locus 2.

Consider the following haplotype probabilities.

 

 

 

 

B b
A pAB pAb m
a Data pab Pa
PB Db

 

 

 

 

 

 

2.3.4.1. (Lewontin’s) D’

The absolute value of the difference between observed and expected

probabilities, under independence, is the same for all four cells, lpABpab‘pAbpaBl-
Let pABpab-pAbpaB = 6, since no cell can be negative the range of possible values
is bounded as a function of the allele frequencies by:
5max '-' minIPbPA- Papa] if 5>0.
6max = minIpApe.-papbl if 6<0.
ID’|=|5l/6max

A disadvantage of the measure is that it is upwardly biased in small

samples. (116)

2.3.4.2. Statistical association using R2

R2: (pABpab‘pAbpaB)2/(pApapob)

For low allele frequencies R2 has more reliable sample properties than

|D’|. R2 measures statistical association and there is a Simple inverse relationship

35

between this measure and the sample Size required to detect association
between susceptibility loci and SNPS has a direct relationship. R2 takes value 1 if

only two haplotypes are present.
2.3.4.3. Estimation of linkage disequilibrium in practice

For designs where unrelated individuals are studied, when considering 2
or more loci at a time, it is not possible to determine which alleles are on one
chromosome in case of heterozygous genotypes. In practice, parental
information is missing (phase is unknown) so one first has to estimate
haplotypes. The most useful information about LD comes from double
homozygotes and one can estimate LD from these alone, but if minor alleles are
uncommon then the estimate of LD may be unstable. It is thus better to get point
and interval estimates of LD parameters by estimating haplotype frequencies.
(116)
2.3.5. Haplotype estimation

When parental genetic information is not available, it is possible to
estimate haplotypes based on the frequencies of alleles of SNPS and their co-
occurrence in the population (in silico methods). Some of the methods used to
estimate haplotypes are maximum likelihood based, implemented via the
expectation-maximization (EM) algorithm. AS the first step, these methods start
from genotypes that their haplotypes are certain. In the next steps, haplotypes of
heterozygous genotype are inferred based on the likelihood methods. For these
genotypes, the most likely combination of haplotype is estimated based on the

frequency of known haplotypes. Generally when there is more than one

36

heterozygous genotype (for example CT-AG, table 2.3) the haplotypes are
ambiguous. In the current work I used PHASE 2.0.2 that utilizes an improved EM

algorithm to estimate haplotypes and construct haplotype pairs. (117)

Table 2.3. An example for haplotype inference based on genotype data.

 

 

SNP1 SNP2 n Genotype n, Haplotype1 n, Haplotype 2
CC AA 6 CA/CA 6, CA 6, CA
CT AA 27 CA/T A 27, CA 27, TA
TT AA 8 TA/T A 8, TA 8, TA
CC AG 85 CA/CG 85, CA 85, CG
CT AG 72 CA/T G 72, CA 72, TG

or or or

TA/ CG 72, TA 72, CG
TT AG 23 TA/T G 23, TA 23, TG
CC GG 485 CG/CG 485, CG 485, CG
CT GG 95 CG/T G 95, CG 95, TG
TT GG 4 TG/T G 4, TG 4, TG
Therefore:
Haplotype P
CA 6+6+27+85=124 0.084
TA 27+8+8+23=66 0.045
CG 85 + 485 + 485 + 95 = 1150 0.784
TG 23+95+4+4=126 0.086
Total 1466

Now the probabilities for possible haplotype pairs of the one with more than
one heterozygous SNP (CT-AG) is calculated:

p1 (CA/TG)=0.084 x0.086
p2 (TA/CG)=0.045 x0.784
p2/p1 = 4.88 -) TA/CG is more likely to be the actual haplotype pair

 

n, frequency; p, probability

37

RESULTS

38

In this chapter three manuscripts, in the order presented, are as follows:

1.

Maternal smoking during pregnancy modifies the effect of IL13 gene
polymorphism on cord serum immunoglobulin E (in preparation)

Raised cord serum immunoglobulin E increases the risk of allergic
sensitization at ages 4 and 10 and asthma at age 10 (Published)

The combined effect of maternal smoking during pregnancy and IL13
gene polymorphism on childhood wheeze phenotypes (submitted)

These manuscripts correspond to the three research questions:

. Is there an interaction between IL13 and maternal smoking during

pregnancy on CS-lgE?
Is CS-lgE related to childhood asthma?
Is there a combined effect of interleukin 13 gene (IL13) polymorphism and

maternal smoking during pregnancy on childhood asthma phenotypes?

39

3.1. Maternal smoking during pregnancy modifies the effect of

IL13 gene polymorphism on cord serum immunoglobulin E

40

Abstract

Interleukin 13 (IL13) has a pivotal role in the pathway of immunoglobulin E
and cord serum immunoglobulin E (CS-lgE) has been suggested to be
associated with allergy later in life. We investigated the association of the IL13
gene polymorphisms on CS-lgE and also the modifying effect of maternal
smoking on this association. In the Isle of Wight birth cohort (UK, 1989-1999),
CS-lgE was measured using the ULTRA EIA® kit and was dichotomized at 0.5
kU/L (n=1358). Five single nucleotide polymorphisms in IL13 were genotyped by
pyrosequencing methods and haplotype analysis was performed. Logistic
regression was used and gene association analysis included 798 children.
Confounders were gender, household cat and dog during pregnancy, position of
child in family and birth weight. Children with the common IL13 haplotype pair
had lower percentage of raised CS-IgE (adjusted odds ratio (AOR) = 0.63, 95%
Cl: 0.41, 0.99). Maternal smoking was not associated with CS-lgE (AOR=1.00,
95% CI: 0.69, 1.46). However, when mothers smoked, the common IL13
haplotype pair was not protective for raised CS-lgE (AOR=0.95, 95% Cl: 0.32,
2.81), but when mothers did not smoke, it was (AOR=0.59, 95% Cl: 0.36, 0.96).
This is the first study that shows an association between IL13 and cord serum
IgE and also a modifying effect of maternal smoking on this association. The
results suggest that the adverse effect of smoking on the immune system may

also be via an epigenetic mechanism.

41

Prevalence of asthma and allergy has increased significantly in the past
three decades. (95, 118) Allergic mechanism has long been attributed to lgE-
mediated reactions and cord serum lgE (CS-lgE) has been suggested to be a
biomarker for atopic diseases later in life (78, 82-88, 119). Although some
researchers disagree with the use of CS-IgE as a biomarker for atopic diseases
(68, 88-93), the value of CS-lgE as a marker for increased susceptibility in later
life on allergy seems important.

There is limited information with regard to the effect of genes and their
interaction with other factors on CS-IgE. Chang et al. reported an association of
cytotoxic T-Iymphocyte antigen-4 (CTLA-4) polymorphism with raised CS-IgE
levels that was limited to girls. (74) Interleukin 13 (IL-13) is an important cytokine
involved in the lgE pathway (35, 72, 105, 106) and yet there is no prior report for
the effect of IL13 gene on CS-IgE.

An important factor on the development of asthma in offspring is maternal
smoking during pregnancy (96-102). A number of studies in children have linked
parental smoking with markers of atopy such as serum immunoglobulin E (lgE)
levels (103, 104). Noakes and colleagues noted that exposure to tobacco smoke
during pregnancy was associated with a significantly higher neonatal lL-13
response to both house dust mite and ovalbumin, indicating that a prenatal
exposure to tobacco may have immunological effects (105). The few studies for
the effect of maternal smoking on CS-lgE did not report consistent results (75-
81). Part of this inconsistency may be due to a different genetic susceptibility of

the studied populations.

42

In the current study we test the hypotheses that IL13 gene polymorphisms
are associated with CS-IgE levels and also explore whether this association is
modified by maternal smoking during pregnancy. The data is from the Isle of
Wight cohort, established in 1989 in order to study the risk factors for atopy and
its development during childhood.

Population and Methods
Study population

The local Research Ethics Committee approved the study of children born
and enrolled (n=1,456) between January 1, 1989, and February 28, 1990, on the
Isle of Wight, United Kingdom. lnforrned consent was obtained and children were
followed up at the ages of 1 (n = 1,167; 80.2 percent), 2 (n = 1,174; 80.6
percent), 4 (n = 1,218; 83.7 percent) and 10 years (n = 1,373; 94.3 percent). The
island is Close to the British mainland, semi-rural, with no heavy industry. The
population is 99 percent Caucasian.

Birth records, exposures during pregnancy and family history

Data was obtained from hospital records on neonatal complications
(preterm delivery, foetal distress, newborn infection, jaundice, low blood sugar,
hypothermia, and congenital abnormalities), antenatal complications (pregnancy
induced hypertension, gestational diabetes, infections needing antibiotic therapy,
intrauterine growth retardation or “miscellaneous”) and method of delivery (non-
instrumental vaginal, instrumental vaginal delivery or caesarean section).

Parents were asked about their age, maternal smoking and pet keeping

during pregnancy, history of asthma, hay fever, and atopic eczema.

43

DNA isolation and IL13 genotyping

Anticoagulated whole blood samples were obtained at the 10-year
interview and stored frozen (n=921). Genomic DNA was isolated from these
samples using QIAamp DNA Blood Kits (Qiagen, Valencia, CA) or the ABI
PRISMTM 6100 Nucleic Acid PrepStation (Applied Biosystems, Foster City, CA).
Polymorphisms in the IL13 gene were examined using the SNPper (120) and
Applied Biosystems (121) databases. Two methods were used for genotyping: 1)
fluorogenic 5’ nuclease chemistry PCR using Assays on Demands kits cycled on
a 7900HT Sequence Detection System (Applied Biosystems, Foster City, CA),
and 2) biotin-streptavidin-based pyrosequencing performed on PSQ-96
instrumentation (Biotage AB, Uppsala, Sweden).

Five SNPS from the IL 13 gene were used in this study, rs1800925 in the 5’
promoter region, rsZO66960 in intron 1, rs1295686 in intron 3, r320541 in exon 4
and rs1295685 in the 3’ untranslated region (3’ UTR) of exon 4. Because IL13 is
a small gene (2.9 kb), only a few SNPS were needed for a reasonable
assessment of genetic associations. Polymorphism selection was based on SNP
location, minor allele frequency, and function. Selected SNPS had minor allele
frequencies 2 19% and were distributed throughout the IL13 gene. The r320541
is a coding variant with the common allele (G) coding for arginine and the minor
allele (A) encoding glutamine at amino acid 144. This SNP was in strong linkage
disequilibrium (LD) with rs1295685 and rs1295686. The former (rs1295685) was
selected as a representative SNP of this block due to the availability of a

commercial assay.

44

Cord serum immunoglobulin E (lgE) determination

Duplicate measurements of cord lgE were made on serum from umbilical
vein blood samples using the ULTRA EIA® kit (Pharrnacia Diagnostics AB,
Uppsala, Sweden), which is designed to measure lgE between 0.2 and 50 kU/L
on 0.1 mL of serum or plasma. It is a solid phase enzyme immunoassay, using
monoclonal anti-human lgE, with a high degree of correlation with the Phadebas
lgE PRIST® method (r=0.99). (122) Cord IgA was measured on samples where
lgE was higher than 0.3 kU/L. Any samples with lgA > 10mg/L were deemed

contaminated by maternal blood and excluded. (122, 123)
Statistical analysis

Each SNP was tested for Hardy-Weinberg equilibrium using Haploview 3.2
software (124). Estimates of LD between SNPS were calculated using 0’ and :2
(116). For SNPS that were in strong LD, the authors took one SNP as the
representative marker for the block. PHASE 2.0.2 was employed to build the
most likely pair of haplotypes (diplotypes) and their probability for each child (117,
125), using the three representative SNPS. In the analysis of individual SNPS and
haplotype pairs, genotypes with low frequencies were combined (126, 127).

Using SAS/STAT® version 9.1, statistical analysis was performed on the
data only from children who had complete information on IL13 genotypes,
exposure to tobacco smoke, and wheeze phenotypes. Chi-square tests were
used to compare the sample used in the analysis with that which was followed up
at age 10. Logistic regression analysis and stratification were conducted to test

the hypotheses. Statistical significance was defined as a p value of s 0.05.

45

Potential confounders (table 3.1.1) were: gender, household cat present
during pregnancy (yes, no).
RESULTS

Data on IL13 genotypes and cord serum immunoglobulin E (CS-lgE) were
available for 798 Children. Compared to the original sample, mothers of Children
who were included in the analysis were younger and less smoked during
pregnancy (p values of 0.03 and 0.0001, respectively, table 3.1.1).

One out of five children was exposed to maternal smoking during
pregnancy (20.5 percent, table 3.1.1). Among covariates, maternal atopy, being
born in fall and first born were risk factors for raised CS-IgE (table 3.1.2).

For IL13 genotypes, all five SNPS were in Hardy-Weinberg equilibrium (p
= 0.82 for rs1800925, p = 0.99 for r32066960, p = 0.58 for rs1295686, p = 0.09
for r320541, and p = 0.36 for rs1295685). Three of the SNPS (rs1295686,
r320541, and rs1295685) were in strong LD (table 3.1.5). Therefore we used only
one SNP of this block for haplotype analysis. Thus haplotypes were inferred
using rs1800925, r82066960 and rs1295685. Due to the limited number of
children homozygous for minor alleles at all SNPS (less than 5 percent, table
3.1.3), minor allele homozygous and heterozygous genotypes were grouped
together. Among haplotype pairs, CCG/CCG had the highest frequency (0.48,
table 3.1.4). The estimated probability for CCG/CCG as the best pair was 1.0 in
99 percent of children and was never less than 0.89 (results not shown). All other
haplotype pairs had frequencies between 0001-0151. In addition, there were

probabilities as low as 0.5 to be identified as the best pair for haplotype pairs

46

other than CCG/CCG. Thus, haplotype pairs other than CCG/CCG were
combined (126, 127).

Crude analysis showed that Children with CCG/CCG haplotype pair had
lower percentage of raised CS-IgE (figure 3.1.1). When the sample was stratified,
in children whose mother did not smoke during pregnancy and also in boys,
CCG/CCG haplotype pair was significantly associated with a lower percentage of
raised CS-lgE (p values of 0.040 and 0.025, respectively, figure 3.1.1). Girls
whose mothers smoked during pregnancy had higher percentage of raised CS-
lgE.

Using logistic regression and controlling for the confounders, we tested
the associations of IL13 haplotype pairs and maternal smoking with CS-lgE
(figure 3.1.2). Children with the common IL13 haplotype pair (CCG/CCG) had
lower relative risk of raised CS-lgE (adjusted odds ratio (AOR) = 0.63, 95% Cl:
0.41, 0.99). Maternal smoking was not associated with CS-IgE (AOR=1.00, 95%
Cl: 0.69, 1.46).

To determine whether maternal smoking modifies the effect of the
common IL13 haplotype pair on CS-IgE, the sample was stratified by maternal
smoking. When mothers smoked during pregnancy, CCG/CCG was not
protective for raised CS-lgE (OR=0.95, 95% Cl: 0.32, 2.81), but when mothers
did not smoke, it was (AOR=0.59, 95% Cl: 0.36, 0.96). We also tested the
multiplicative interaction between CCG/CCG and maternal smoking on raised
CS-IgE in a logistic regression model. This interaction was not significant;

however, the interaction term between gender, maternal smoking during

47

pregnancy and CCG/CCG haplotype pair was Significant (p < 0.05). To
demonstrate this interaction we looked at the effect of gender when the other two
factors were fixed. The odds ratio for raised CS-lgE in girls whose mothers
smoked during pregnancy and had CCG/CCG haplotype pair compared to boys
whose mothers smoked during pregnancy and had CCG/CCG haplotype pair
was 7.92 (results are not shown).

DISCUSSION

Using the data from the Isle of Wight birth cohort, this study investigated
the effect of haplotype pairs of the IL13 gene, maternal smoking during
pregnancy and gender on cord serum immunoglobulin E (CS-lgE). IL13 gene
polymorphisms were associated with CS-lgE levels. Maternal smoking during
pregnancy was not related to CS-lgE levels, but its combination with the common
IL13 haplotype pair was associated with raised CS-lgE, especially in girIs. To the
best of our knowledge this is the first investigation for the association between
IL13 and CS-lgE.

For this study, the information was available from a subset of children who
had a CS-lgE measurement and who agreed to provide blood for genotyping.
These children appeared to have younger mothers and also less exposure to
maternal smoking in comparison to all Children who had information on CS-IgE.
As the distribution of the outcome variable, CS-IgE, was not different in the
original sample and the sample used in the analysis, estimates were not affected
by selection process. (128) With regard to genotypes, presence of a selection

bias could result in a violation of the Hardy-Weinberg equilibrium. The genotypes

48

of the five SNPS used in the analysis were in Hardy-Weinberg equilibrium and
their allele frequencies were comparable to the equivalent SNPS in other
Caucasian populations (72, 73). Hence, with regard to the SNPS a selection bias
is unlikely.

Contamination of cord blood with maternal blood was assessed using lgA
as a biomarker. This led to the exclusion of 86 infants from further analysis. (129)
(130) Hence, a potential misclassification of cord blood lgE is unlikely. To
dichotomize CS-IgE levels, cut-off points between 0.5-1.3 kU/L have been used
in prior studies. (77, 84, 86, 131) We chose to use the cut off point of 0.5 kU/L
CS-lgE as it has been Shown in previous analyses of this birth cohort as an
optimal level. (129) (119).

With regard to genetic analysis, information on individual SNPS was used
and the most likely pairs of haplotypes were estimated from genotype data. It
would have been possible to use information on each individual SNP to
investigate gene-environment interactions. However using haplotype pairs has
some advantages over individual SNP analysis. The issue of “multiple testing” is
avoided as all genetic markers were compiled in haplotype pairs. Additionally, it
has been suggested, Specifically for IL13 (132), that haplotype analysis could
confer more information than individual marker analysis (132-134). Haplotype
pair analysis may misclassify genotypes when parents’ genetic information is not
available (ambiguous phase). However, the probability of having CCG/CCG, the
major haplotype pair, was 1.00 in 430 out of 435 children with this genotype (the

probability for the other 5 children was 0.89). Thus, CCG/CCG haplotype pair

49

was inferred with high certainty. The distribution of the data, with respect to minor
haplotypes, did not allow testing for their interactions with tobacco smoke

exposure. However, when children with a minor haplotype at both loci (frequency
= 0.082, table 3.1.4) were removed from the analysis, we obtained similar results.

Previous studies have shown an association between IL13 polymorphisms
and serum lgE. (72, 108, 135, 136) However, there is no report for the effect of
IL13 polymorphisms on cord serum lgE. With regard to the effect of maternal
smoking on CS-lgE, results from previous studies are not consistent. While some
reported maternal smoking as a risk for higher levels of CS-IgE;(76, 77, 81, 137),
others did not find such an association (75, 78, 80).

A different prevalence of allergic disorders in boys and girls has long been
recognized. (138) In the only publication on the association of genes and cord
serum lgE, Chang et al reported a link between cytotoxic T-lymphocyte antigen-4
(CTLA-4) polymorphism and cord serum lgE levels that was limited to girls. (74)
This gender difference has been also shown for the effect of maternal smoking
on pulmonary function tests. (139, 140) Some researchers have explained the
gender discrepancy based on the effects of sex hormones. (141, 142) There is
evidence that sex steroids and substances that mimic their physiological effect
can increase allergic manifestations. (20, 143) There is evidence that fetal
gender has an influence on hormone profiles. For example, placenta produces
more hCG levels when the fetus is female. (144-147) The latter findings suggest
a difference in placental function in boys and girls. The effect of gender on the

relationship of IL13 polymorphisms and their interaction with maternal smoking

50

 

on CS-lgE may result from a different activity of placenta in girls and boys and
also various amounts of exposures due to placental activity. It is plausible to
consider that maternal smoking, independent of fetal gender may also interplay
with IL13 polymorphisms to increase CS-lgE. Liu et al. reported that the
association of IL13 polymorphisms and serum lgE is modified by exposure to
tobacco smoke. (148)

It has been shown that exposure to tobacco smoke can activate cytokine
genes. (149) As mentioned, there is a disagreement between previous studies
that investigated the effects of maternal smoking on CS-lgE. (75-78, 80, 81, 137)
If we assume that the genetic make up of children studied in previous reports
were different, the observed disagreement for the effect of maternal smoking on
CS-lgE may be due to an interaction between exposure to tobacco smoke and
genes.

In summary, in a sub-sample of the Isle of Wight cohort, the combined
effect of exposure to tobacco smoke during pregnancy and the common
haplotype pair of the IL13 gene resulted in an increased relative risk of raised
CS-lgE, especially in girls. These results highlight the importance of the interplay
of prenatal exposures and genetic predisposition as a mechanism for allergy.
Given that the genetic make up is different across studies, the current report
suggests that negative reports for the effect of exposure to tobacco smoke on
CS-lgE (75, 78, 80) may be dependent on genetic variations. Additionally the
observed role of gender for the effect of IL13 makes us propose future studies to

investigate the expression of IL13 in boys and girls and by various exposures.

51

These studies may Shed light on possible mechanisms of gender discrepancy in

allergic disorders.

52

Table 3.1.1 Comparison of children with a follow-up at age 10 and the subset
used in the analysis

 

 

Numbers at Numbers used in 2
birth (%) the analysis (%) x pvalue
Variable Total 1 ,358 798
2 34 389 (28.6) 211 (26.4)
Maternal age 224. <34 801 (59.0) 488 (61.1) 003
(year) <24 145 (10.7) 93 (11.6) '
Missing 23 (1.7) 6 (0.7)
Maternal 1 438 (32.2) 256 (32.1)
atopy 0 905 (66.6) 538 (67.4) 0.95
Missing 15 (1 .1) 4 (0.5)
Paternal 1 352 (25.9) 215 (26.9)
atopy 0 983 (72.4) 575 (72.1) 0.51
Missing 23 (1.7) 8 (1.0)
Maternal 1 334 (24.6) 164(20.5)
smoking 0 1012(74.5) 631 (79.1) 0,0001
Missing 12 (0.9) 3 (0.4)
Household Yes 406 (29.9) 253 (31.7)
d°9 No 935 (68.8) 540 (67.7) 0.41
Missing 17 (1.2) 5 (0.6)
Household Yes 434 (32.0) 267 (33.5)
cat No 907 (66.8) 526 (65.9) 0.45
Missing 17 (1.2) 5 (0.6)
Season of Fall 296 (21.8) 171 (21.4)
birth Spring 315 (23.2) 187 (23.4)
Summer 305 (22.5) 173 (21.7) 0'76
Winter 442 (32.5) 267 (33.5)

53

Table 3.1.1 (cont’d)

 

Gender

Position of

child in family

Antenatal

complications

Non-
instrumental
delivery

neonatal

complications

Cord lgE

Girl
Boy
Missing
2

1
Missing

Yes
No

No
Yes
Missing

Yes
No

2 0.5 kU/dl
< 0.5 kU/dl

661(48.7)
696 (51.2)
1 (0.1)

636 (46.8)
446 (32.8)
276 (20.3)

132 (9.7)
1226 (90.3)

174 (12.8)
924 (68.0)
260 (19.1)

187 (13.8)
1171 (86.2)

174 (12.8)
1184 (87.2)

394 (49.4)
404 (50.6)
0 (0.0)

432 (54.1)
289 (36.2)
77 (9.6)

81 (10.1)
717 (89.8)

100 (12.5)
543 (68.0)
155 (19.4)

110 (13.8)
688 (86.2)

96 (12.0)
702 (88.0)

0.56

0.26

0.53

0.69

0.98

0.30

 

54

36236856 _mzcomcoo

.EEofian .mezm vooE So. .8656.— .co_8£c_ 5326: .86me .308 $62.96 E965 co Ewcoofi
5:69.39 £39m 65.65655
new 3865 6:05:26 msnooc 30:69.5 £96nt _mcosfimom £08556? 30:9: 65:83 no 86:00:.
cozoom cmocmmomo ocm 302.66 .659? _EcoEEfiE .m ”.252, :_ Eom .N 6mm mam _mEQms. é

86:05 8:966”.

 

55

 

t6 8.? 66 6: 3:5 8.8 $2 :8: 8.2 8:88:38 .988:
86 8.9 .86 8.: 86:8: .89 88:8: 8.3 n82.8 88885.82
S6 was .86 86 3:6: .86 8F :66: 6:2 8:28:38 389%
86 :6 .Be 86 835.: 8.8 885: .88 58 .262 889, Em
86 «3 .86 6 Z A: :8: 383.8 82 :89 .8. 8 6m
86 86 .86 666 8:8: 9:64 38:65 8.8 moseso E E8
86 8. F. 86 8.: 88: side 8635 .8. :4 N688 5 Eom
666 86 .86 8: 36:8: .88 C888: .88 was 5 Eom
8.6 N: .86 mm; 82:8: .32.» 8F :8: SEN :8 6.888...
26 63 66 8.? 5:8: .88 62 :85 38.8 68 6.288:
666 63 .86 66.? 828: $6.36 at :8: 86m 6:658 .6598:
$6 84 .86 N3 8:8: as: F. 69 :68: .88 >88 .8686
~66 8.6 .Be 63 5:6: .864 at :8: .64... 8 .32.... .8882
86 6.3 .86 86 :88: .i. K 8:8: $6.8 XV .ewm .68 86522
86 :3 .86 «6 8:3: .88 38:66: .88 8 n .28 8.5522
8.9 n. _o s. 8 «.0 8.6 n 66 v 98960

 

md. E23 200

 

 

w c__:no_moc:EE_ E263 Eco :0 86:98 Lo: 6289 9 2E o. .8sz wave nominee: .N. to 636...

Table 3.1.3. Single nucleotide polymorphisms (SNPS) in the IL13 gene
for this analysis

 

 

SNP Position (bp) Location Genotype Frequency (%)
rs1800925 132,020,708 Promoter CC 577 (63.6)
CT 295 (32.5)
TT 35 (3.9)
Total 907 (100.0)
r32066960 132,022,334 Intron 1 CC 729 (81.5)
AC 157 (17.6)
AA 8 (0.9)
Total 894 (100.0)
rs1295686 132,023,742 Intron 3 CC 483 (64.6)
CT 240 (32.1)
TT 25 (3.3)
Total 748 (100.0)
r320541 132,023,863 Exon 4 66 73 (53.7)
AG 59 (43.4)
AA 4 (2.9)
Total 136 (100.0)
rs1295685 132,024,344 Exon 4 GG 584 (64.5)
AG 280 (30.9)
AA 41 (4.5)
Total 905 (100.0)

 

Gene annotation is based on SNPper (120).
IL13: interleukin 13 gene; n, number of samples used in genetic studies;
bp, basepairs.

56

Table 3.1.4. Frequency of haplotypes and haplotype pairs for three Single
nucleotide polymorphisms (rs1800925, r52066960, rs1295685) of the
interleukin 13 gene, inferred from genotype data of 907 children

 

 

Frequency (SE)
Haplotype CCG 0.683 (0.004)
CCA 0.045 (0.002)
CTG 0.067 (0.003)
CTA 0.108 (0.003)
ACG 0.041 (0.003)
ACA 0.029 (0.002)
ATG 0.008 (0.002)
ATA 0.017 (0.003)
Haplotype pair(S) CCG/CCG 0.480
CCG/minor haplotypes* 0.438

minor haplotypes/minor haplotypest 0.082

 

PHASE 2.0.2. was used for this analysis (117, 125).

* This group consists of, with a frequency of:

CCG/ACA, 0.036; CCG/ACG, 0.055; CCG/ATA, 0.037; CCG/ATG, 0.019;
CCG/CCA, 0.054; CCG/CTA, 0.151; CCG/CTG, 0.086.

1' This group consists of, with a frequency of:

ACA/ATA, 0.001; ACG/ACA, 0.003; ACG/ACG, 0.001; ACG/ATA, 0.002;
ATG/ATA, 0.001; CCA/ACA, 0.004; CCNCCA, 0.002; CCA/CTA, 0.014;
CTA/ACA, 0.014; CTA/ATA, 0.002; CTA/CTA, 0.007; CTG/ATA, 0.004;
CTG/CTA, 0.020; CTG/CTG, 0.004.

57

 

Table 3.1.5. Using the Haploview program (124),
three Single nucleotide polymorphisms of the
interleukin 13 gene appeared to be in a block.

 

Pair of SNPS

,2

o' (95% Cl)

 

rs1800925 — rs2066960:
rs1800925 — rs1295686:
rs1800925 — rs20541:

rs1800925 — rs1295685:
r62066960 — rs1295686:
32066960 — rs20541:

r52066960 — rs1295685:
31295686 — rs20541:

rs1295686 - rs1295685:
rs20541 — rs1295685:

0.01
0.31
0.28
0.28
0.06
0.07
0.06
0.78
0.90
0.85

0.14 (0.04, 0.24)
0.56 (0.50, 0.62)
0.63 (0.46, 0.75)
0.54 (0.48, 0.59)
0.37 (0.27, 0.47)
0.37 (0.04, 0.71)
0.38 (0.28, 0.47)
1.00 (0.83, 1.00)
0.96 (0.93, 0.98)
1.00 (0.94, 1.00)

 

D’and r2 are pair wise linkage disequilibrium
determinants (116); CI, confidence interval.

58

.63 .2 S 0.03:8 Nod ww<Id 9:0: 68890.
new oomooomm. .mmmoomvm. .0c0m 9 5030.005 0o mEmEEoE>_oa 0260.03: 0650 00.5 0o Emu 0qboc0m E0: 00:00:. ._.

000098 .wa+ 60890 8: 3:0.

05:33:50 :05 new 289 55305

@550 mcioEm .0599: 2 05890 g0oc0m E 00:30:» 0:0 Can "5 0.9500 .38 :_ 59050. new 0005009
+w0aboc0m 2.: *0 09.05 oz: 05 .8 33x mdAv m c__:no_mo::EE_ 820m Boo c003. 0o 090E020; fifm 059...

 
 
 
 

 

    

 

 

   

  

 

 

 

 

 

 

 

 

 

 

 

 
 
 

 

 

 

 

 

 

 
 

 

 

€9.36 €04.30 1.9.250 €0.50 .50 :8 :2? .20. .58 o
I .l I
ll m
I i a
my
8... u a 8.0 u q u q B... u a «a... u q «a... u q 8... u a 3... u a B... u a an
£050 I 000500 B

 

 

 

 

 

mm

59

 

 

2.5

 

 

 

 

 

 

.1, l i T

1) 2) 3) 4)

 

Odds ratio for cord serum lgE > 0.5 kU/dl
0|

 

 

Figure 3.1.2. The association between CCG/CCG haplotype pair and smoking
with raised cord serum immunoglobulin E (CS-lgE >0.5 kU/dl).

1) CCG/CCG versus other haplotype pairs

2) Mother smoked during pregnancy

3) CCG/CCG, mother did not smoke during pregnancy

4) CCG/CCG, mother smoked during pregnancy

60

3.2. Raised cord serum immunoglobulin E increases the risk of

allergic sensitisation at age 4 and 10 and asthma at age 10

61

Background: Evidence suggests that elevated cord serum lgE (CS-lgE) is a risk
factor for allergic sensitisation. However, whether CS-lgE is a risk for asthma is
controversial.

Objective: To investigate the association between CS-lgE levels and allergic
sensitisation at 4 and 10 years and asthma at 1-2, 4 and 10.

Methods: CS-lgE was available for 1358 of 1456 children born between 1989
and 1990. The cohort was evaluated for allergic diseases at ages 1, 2, 4 and 10
years. Skin prick tests for six aero-allergens were performed on 981 children at
age 4 and 1036 at age 10. Asthma was defined based on physician's diagnosis.
Using logistic regression analysis we estimated the risk of asthma and allergic
sensitisation for elevated levels of CS-lgE (20.5 kU/L).

Results: At ages 4 and 10 years, 16.5% and 27.0% of children, respectively,
showed allergic sensitisation. The risk of allergic sensitisation was significantly
associated with elevated CS-lgE at ages 4 (OR: 2.29) and 10 years (OR: 1.73).
The prevalence of asthma was 10.3% at age 1-2, 15.2% at age 4 years and
12.8% at age 10. CS-lgE was not associated with asthma at age 1-2 and 4, but
showed an increased relative risk at age 10 (OR: 1.66, 95%Cl 1.05-2.62). The
association was stronger in children who did not develop allergic sensitisation at

age 4 or 10 (OR: 3.35, 95%Cl 1.41-7.93).

Conclusions: Elevated cord serum lgE is a risk factor for allergic sensitisation at
ages 4 and 10 years. This is the second study suggesting that CS-lgE is also a
risk factor for asthma at age 10, probably related to late onset of asthma. This

association is not necessarily mediated by allergic sensitisation.

62

Introduction

There is evidence that factors early in life have major effects on the later
development of allergy (atopy). (78, 82-88) A correlation between serum lgE
during the first year of life and atopic diseases in the first 2 years of life was first
reported in 1975 by Orgel et al., which led them to propose that cord serum lgE
(CS-lgE) might be used as a predictor for allergic diseases. (82) Subsequent
studies examined the use of CS-lgE as a biomarker for atopic diseases, but
results were conflicting. (68, 83, 88-93) A limitation of these studies is that only
early childhood manifestations (up to 5 years) were investigated,(78, 82, 90, 94)
with the exception of a Swedish study that followed newborns to the age of 11.
(88) In addition, almost all of the previous studies used total CS-lgE (Nambu et
al. investigated CS-lgE specific to a particular allergen (83)). Comparisons
among studies are further complicated by their use of different allergic outcomes.

The Isle of Wight cohort was established in 1989 in order to study the risk
factors for atopy and its development during childhood. The first examination of
CS-lgE in this cohort revealed it to be an insensitive predictor of asthma,
eczema, and hay fever in the first year of life. (122) A subsequent study of 4 year
olds by Tariq et al. showed that elevated CS-lgE increases the risk of
aeroallergen sensitisation without increasing respiratory allergic symptoms. (129)
The purpose of the current study is to investigate the association between CS-
IgE levels and both allergic sensitisation and asthma in the Isle of Wight cohort

up to the age of 10 years.

63

POPULATION AND METHODS
Study population

After approval by the Local Research Ethics Committee, 1456 of children
born between January 1989 and February 1990 on the Isle of Wight, England,
were recruited for the cohort. lnforrned written parental consent was obtained for
all participants. To measure cord serum lgE, samples were obtained from 1358
newborns.

Questionnaire and examination

In the first survey, at the time of the child's birth, parents were asked about
their age, maternal smoking during pregnancy, and history of asthma, hay fever,
and atopic eczema. At 1 and 2 years of age, a questionnaire was administered
seeking information on symptoms that are suggestive of allergic disease in
children.

At 4 years of age, all children and their parents were invited to attend the
allergy clinic. (10) A total of 981 children (67% of the original cohort) attended the
clinic, a questionnaire was completed, and a physical examination was
conducted. A follow up questionnaire (postal or telephone) was used to gather
information for an additional 237 children (n=1218; 84% of the original cohort)
who had not attended the clinic. Some families did not participate in the study
because they moved from the Isle of Wight, while others declined to have their
child skin prick tested. Questionnaire information was collected on the presence
of allergic diseases. Birth order data and details of environmental factors such as

parental smoking, exposure to pets and housing conditions, were updated.

64

At age 10 years, information was updated and physical examination was
conducted during a visit to the Research Center (n=1043). Detailed interviewer-
administered questionnaires were completed with a parent of each child
regarding asthma and allergy development. ISAAC written questionnaires were
used to assess respiratory, nasal, and dermatological symptoms. (16) When a
visit was not possible, a telephone questionnaire was completed with a parent, or
a modified postal questionnaire was sent for the parent to complete and return
(n=330)

Extract of birth records

Data was obtained from hospital records on gestational age, birth weight,
and neonatal complications (preterm delivery, foetal distress, newborn infection,
jaundice, low blood sugar, hypothermia, congenital abnormalities, etc.).

Cord serum lgE determination

Duplicate measurements of cord lgE were made on serum from umbilical
vein blood samples using the ULTRA EIA® kit (Phannacia Diagnostics AB,
Uppsala, Sweden), which is designed to measure lgE between 0.2 and 50 kU/L
on 0.1 mL of serum or plasma. It is a solid phase enzyme immunoassay, using
monoclonal anti-human lgE, with a high degree of correlation with the Phadebas
lgE PRIST® method (r=0.99). (122) Cord lgA was measured on samples where
lgE was higher than 0.3 kU/L. Any samples with lgA>10mg/L were deemed
contaminated by maternal blood and excluded. The method of cord lgE

estimation was described in greater detail in previous reports. (122, 123)

65

Diagnostic criteria for asthma

The study physicians obtained information regarding the presence of
allergic symptoms. For asthma, information included the presence and frequency
of cough and wheezing, physician-diagnosed asthma, and treatment given. At
ages 1, 2 and 4 the minimum criteria for the diagnosis of asthma included three
or more separate episodes of wheeze in the previous 12 months, each lasting at
least three days. At age 10, asthma was defined as ever having a diagnosis of
asthma, in addition to wheeze in the last 12 months.
Skin prick tests

Skin prick tests (SPT) were carried out on children at age 4 years (n=981)
and 10 years (n=1036) using a standard battery of aeroallergens (house dust
mite [Dermatophagoides pteronyssinus], grass pollen mix, cat and dog epithelia,
Altemaria altemata, and Cladosporium herbarum). Histamine (0.1%) in
phosphate-buffered saline and physiological saline were used as positive and
negative controls, respectively. Extracts were from Biodiagnostics, Hamburg,
Germany at age 4 and ALK-Abello, Denmark at age 10. Children were advised
not to take antihistamines for 72 hours before the clinic appointment.
Commercially available lancets (Medipoint, |nc., Mineola, NY) were used to prick
the epidermis through the allergen extract drops. The tests were read after 15
minutes and a mean wheal diameter (the sum of the longest diameter and the
diameter diagonal to it divided by two) of at least 3 mm greater than the negative

control was taken as positive. Surrounding erythema was ignored. A child was

66

regarded as sensitised when the reaction to at least one allergen was positive
and non-sensitized when all recorded reactions were negative.
Statistical analysis

We used likelihood ratios (=sensitivity/ (1-specificity)) to compare cut-off
points for CS-lgE. CS-lgE levels were then dichotomised into less than 0.5 kU/L
and greater or equal to 0.5 kU/L. Asthma at age 1-2 years was regarded as
positive when at least one positive record was available at either age 1 or 2
years, and negative when there was at least one negative record and no positive
record. SPT and asthma at different ages were treated as dichotomous outcome
variables and separate logistic regression models were run. Potential
confounders were: seasons of birth (spring: March-May; summer: June-August;
fall: September-November; winter: December-February); gender; birth weight
(categorized as: <15009, 1500-25009, 2500-40009, 240009): neonatal
complications (yes, no); maternal age (categorized as: 16-23, 24-33, 34 years
and older); history of maternal and paternal atopy (asthma, hay fever, atopic
eczema); maternal smoking during pregnancy (yes, no); and having a pet at birth
(yes, no).

To test the effect of CS-lgE levels on sensitisation and asthma we used
logistic regression, adjusting for the above-mentioned potential confounding
factors (LOGISTIC procedure). We estimated adjusted odds ratios of CS-lgE for
SPT and asthma at different ages. For asthma, we additionally stratified data for
positive SPT. We estimated adjusted odds ratios of SPT and asthma at different

ages for the two levels of CS-lgE. All statistical analyses were performed using

67

the SAS/STAT® program. (112)
RESULTS

Cord serum lgE (CS-lgE) was determined for 1358 children (93.3% of the
original cohort). The proportion of newborns with increased CS-lgE (20.5 kU/L)
for various characteristics is shown in Table 3.2.1. A maternal history of atopic
disorders was a significant risk factor for increased CS-lgE. The number of
children who had complete information on CS-lgE, SPT and all confounders was
862 at age 4 and 900 at age 10 (63.5% and 66.3% of the sample with available
CS-lgE; table 3.2.2). Regarding asthma, children who had complete information
on this outcome, as well as on CS-lgE and confounders, were 1194 at age 1-2,
1068 at age 4, and 1191 at age 10 (ranging between 86.9% and 87.3% of the
sample of 1358; table 3.2.2). The proportion of children with elevated CS-lgE
levels (20.5 kU/L) in the samples for different outcomes at different ages did not
vary substantially (12.3-13.3%, table 3.2.2).

Likelihood ratios for different cut-off points of CS-lgE indicated the level of
risk for allergic sensitisation and asthma at age 4 and 10 (table 3.2.3). The
Likelihood Ratio (LR) is the likelihood that a specific level of a test result would
be expected in a child with a disorder compared to the likelihood that that same
result would be expected in a child without the disorder. (150) CS-lgE levels
above the detection limit increased the likelihood of the development of allergic
sensitisation (LR>1), but not of asthma. CS-lgE levels of 0.5 kU/L showed higher
likelihoods ratios in three of the four outcomes and thus provided more

confidence in abnormal test results.

68

A positive SPT was evident in 16.5% (142/862) of the children at age 4
and in 27.0% (243/900) at age 10 (table 3.2.4). CS-lgE was a risk factor for a
positive SPT both at age 4 and at age 10, however, the association was stronger
at age 4 years (table 3.2.4, figure 3.2.1). Regarding specific SPTs, the
prevalence increased for most allergens. However, the association (OR) between
CS-lgE and allergic sensitisation decreased between 4 and 10 years (table
3.2.4).

At age 1-2, physician diagnosed asthma was found in 10.3% of the
children (123/1194, table 3.2.5). The prevalence of asthma at age 4 was 15.2%
(162/1068) and at age 10 was 12.8% (152/1191). Logistic regressions were used
to test the associations of CS-lgE with asthma at age 4 and 10. In addition, to
assess whether these relationships were mediated by sensitisation, we stratified
asthma at age 4 by SPT positivity at age 4 and asthma at age 10 by SPT
positivity at age 4 or 10. Table 3.2.5 shows that cord serum lgE is not associated
with asthma at age 4, irrespective of atopic status. The association of CS-lgE
with asthma at age 10 showed an OR of 1.66 (95% Cl 1.05 to 2.62), which is
statistically significant. When stratifying by SPT results, the association lost its
significance in children who had a positive SPT at age 4 or 10 (OR=0.83).
However, the majority of children had a negative SPT at both ages. In these
children the association between CS-lgE and asthma at age 10 became stronger
(OR=3.35, 95% Cl 1.41 to 7.93). Figure 3.2.1 shows that with increasing age of
diagnosis the relative risk (OR) of CS-lgE for asthma increases.

The dynamics of the development of asthma are shown in figure 3.2.2. Of

69

the 138 children who were diagnosed with asthma at age 10, only 44 had asthma
at age 1-2. Fifty-one children were first observed to have asthma at age 10 (37%
of the children with asthma at age 10). For each subgroup in figure 3.2.2, the
percentage of children with elevated cord serum lgE is shown. In children with
late onset of asthma (developing after 4 and before 10 years), the proportion is
substantially higher (20%), compared, for instance, to the subgroup with early
onset of asthma at age 1-2 years (11.2%, figure 3.2.2).
DISCUSSION

Our analysis revealed that elevated cord serum lgE (20.5 kU/L) is a risk
factor for allergic sensitisation at ages 4 and 10 years. Elevated cord serum lgE
(CS-lgE) was also shown to be a risk factor for asthma at age 10 years, but not
for asthma at age 1-2 or at 4 years. To the best of our knowledge this is the
second study of the association between CS-lgE and asthma at a later age in
childhood. (77, 82, 131) Studies over the past three decades conducted on the
association between CS-lgE and asthma in childhood are not in agreement. (78,
84, 129, 151, 152) Most of these studies involved a small number of children and
followed them to early childhood (up to 5 years). (77, 151, 152) Additionally,
some studies were based only on a clinical assessment of atopy, not on a
determination of allergic sensitisation. (77, 82, 131) Our results corroborate prior
studies which found that CS-IgE is a risk factor for allergic sensitisation in early
childhood and extend this observation, for the first time, to the age of 10 years.
Our findings also support the report of Croner et al. (153) that elevated CS-lgE

increases the incidence of asthma, cumulated during childhood up to the age of

70

11 years (four investigations at 1.5, 3, 7, and 11 years). (86, 88, 153)

The analyses are based on samples of those children with complete
information on the outcome (sensitisation or asthma), exposure (CS-lgE) and
potential confounders. Comparisons of the initial sample that had CS-lgE
measurements (n=1358) and the samples used in the analyses indicate that
selection bias is unlikely (table 3.2.2).

Contamination of cord blood with maternal blood was assessed using lgA as
a biomarker. This led to the exclusion of 86 infants from further analysis. (129)
(130) Hence, a potential misclassification of cord blood lgE is unlikely. To
dichotomize CS-lgE levels, cut-off points between 0.5-1.3 kU/L have been used
in prior studies. (77, 84, 86, 131 ) We chose to use the cut off point of 0.5 kU/L
CS-IgE for two reasons. First, it has been shown in previous analyses of this birth
cohort as an optimal level. (129) Second, likelihood ratios indicated that the
certainty of atopic development is increased for CS-lgE of >0.5 kU/L (LR > 1.58
for three of four outcomes). Higher levels of CS-lgE were less reliable because of
the small number of children (e.g. 21.0 kU/L and asthma at age 10: n=9). To
assess allergic sensitisation, we used SPT, accepted as an objective marker for
evaluating allergic sensitisation. (154, 155) The use of both negative and positive
controls diminished the possibility of false results in this study.

lgE production is downstream in a pathway that starts with the

presentation of antigens (allergens) to the immune system. Antigen presentation
to the immune system may result in the preferential activation of Th2 cells which

are programmed to secrete lL-4 and IL-13 and to induce lgE production. (156-

71

158) An important question is whether the association of cord lgE with asthma is
due to the relation of cord lgE with early sensitisation. Our results indicate that
the direct effect of cord serum lgE disappears in children who developed a
sensitisation (SPT at age 4 or 10 years, table 3.2.5). Since CS-lgE is also a
significant predictor of SPT (table 3.2.4), CS-lgE is likely to have an indirect
effect on asthma: CS—lgE 9 SPT positivity -) asthma. However, sensitisation is
not a necessary link in the chain of reactions. The reason is that in children who
did not develop a sensitisation, representing the majority, a strong direct effect
between CS-lgE and asthma became obvious (OR=3.35, table 3.2.5). In addition,
we investigated whether considering only early sensitisation (at age 4) changed
the direct and indirect effect of CS-lgE on asthma. The results did not differ
substantially from those when SPT at age 4 and 10 were combined.

Pet exposure at birth was not associated with cord serum lgE levels (table
3.2.1) and did not explain the occurrence of allergic sensitisation and asthma. In
addition, we examined whether additional control of pet exposure at age four
changed the results. This was not the case. We also controlled for the number of
children in the home which did not change our findings. Lastly, we also stratified
for pet at birth and at age four as well as for the number of children at home. We
did not identify any combined effect of these factors with CS-IgE.

It is widely accepted that dramatic changes in immune responses occur

during pregnancy and foetal markers of immune response are detected as early
as 10-22 weeks gestation. (60) To prevent the termination of pregnancy, the

foeto-matemal immune system is shifted in a Th2 direction. (62, 63) This Th2

72

shift begins with the secretion of human chorionic gonadotropin, which in turn
stimulates progesterone production. Progesterone then promotes the secretion of
Th2 cytokines over the secretion of Th1 cytokines. (60) Although cytokine
production may be variable in the early stage of development, the Th1 and Th2
subsets become stable after repeated stimulation. (156-158) There are two
alternative concepts regarding the time when sensitisation and T-cell stability
occurs: the prenatal priming hypothesis considers that foetal life immunologic
events are more important in the development of allergy in childhood, and the
hygiene hypothesis, which emphasises that the insults happen after birth.

Given that CS-lgE is a biomarker of prenatal immunologic conditions, our
findings favour the prenatal pn'ming hypothesis. CS-lgE was associated with
overall SPT at ages four and 10 (table 3.2.4). The presence of a significant
adjusted OR for the overall SPT and specific SPTs both at age 4 and 6 years
later, supports this idea. There is a non-significant decline in ORs and
percentage of elevated CS-IgE at age 10, which may be a reflection of the
weaker effect of postnatal life events as opposed to prenatal life events (table
3.2.4, figure 3.2.1).

For elevated CS-lgE levels, we observed a stronger association between
maternal history of atopy in comparison to paternal history of atopy (table 3.2.1).
This observation is in agreement with previous findings and other evidence that
maternal immune responses may influence allergic responses in offspring. (159-
162)

For asthma at age 10, CS-lgE gained a prognostic value (OR=1.66). In the

73

one prior investigation with a comparable follow-up time after birth, Croner et al.
showed a 5-fold increase in the cumulated incidence of asthma in children at 11
years who had cord serum lgE levels of 0.9 kU/L and higher. (153)

At ages 1-2 and 4 years the minimum criteria for the diagnosis of asthma
included three or more separate episodes of wheeze in the previous 12 months,
each lasting at least three days. At age 10, the definition of asthma was at least
one episode of wheezing in the last 12 months in addition to being diagnosed
with asthma by a physician at least once. Different criteria for asthma diagnosis
at age 4 and 10 may have led to misclassification. This misclassification, should
it exist, would be non-differential. A previous diagnosis of asthma was less likely
to have biased asthma diagnosis at age 10 because fewer children were
diagnosed as asthmatics at age 10 (12.8%) in comparison to age four (15.2%). It
is unlikely that the level of CS-IgE may have affected asthma diagnosis at age 10
but not at age 4, as this information was not available to the field workers who
collected information on the presence of wheeze and asthma diagnosis.

As mentioned, there is evidence that early life events may influence atopy
and asthma later in life. But how could CS-lgE information predict asthma at age
10 years but not at an earlier stage of life? One possible explanation is that in
early childhood diagnosis of asthma is based on transient symptoms that tend to
disappear later in life. Kurukulaaratchy et al. showed that approximately 2/3 of
children who had wheezing in the first four years of life lost their wheezing by age
10. (163) Additionally, it has been reported that most infants who wheeze are not

necessarily atopic and that many lose their wheezing as they grow older. (164)

74

Assuming this kind of sorting process, the diagnosis of asthma becomes more
reliable for children with persistent wheezing. However, our data does not
support this explanation, as the percentage of elevated CS-lgE in those who
persistently had asthma at ages 1-2, 4 and 10 years (12.5%) was not different
from those who never had asthma (12.3%, figure 3.2.2).

Another possible explanation is that the late onset of asthma is related to
CS-lgE. In this case, we would expect a higher proportion of children with
elevated CS-lgE in the late onset group. In our study, 51 children (figure 3.2.2)
had asthma at age 10 but not earlier; of this group, only a few had wheezing
before age 10 (6% at age 1-2 anle% at age 4). However, 20% of these children
had elevated CS-lgE, which is the highest proportion of any of the groups (figure
3.2.2). The next highest proportion of CS-lgE (16.6%) was found in children who
had asthma at age 4 and again at age 10 years. Hence, our results support the
late-onset explanation over the sorting process explanation.

As seen in figure 3.2.1, the trend for ORs for asthma increases until it
reaches significance at age 10. Additionally, the percentage of asthmatics among
those who had elevated CS-lgE was 9.0% at age 1-2, 14.8% at age 4 and 18.8%
at age 10 years. These observations could be regarded as the gradual
development of asthma in those who had elevated CS-lgE and hence support
the importance of CS-IgE as a biomarker of feotal life events in the development
of asthma later in life.

In conclusion, elevated cord lgE measured at birth increases sensitisation to

aeroallergens by the ages of 4 and 10 years as well as asthma at 10 years, but

75

not asthma at younger ages. These findings support the importance of prenatal
conditioning in the development of allergy later in life. The lack of significance of
CS-lgE as a risk factor for asthma early in life may be due to the gradual
development of asthma as the cohort aged and the existence of a late-onset
variety of asthma. Our findings encourage investigations which would lead to an
increased understanding of the probable immunologic mechanisms that are
responsible for the effect of foetal life events on childhood allergy. Additionally,
we recommend long-term assessments because of the possibility of late-onset

childhood allergy.

76

 

 

Table 3.2.1 Characteristics of children with an available CS-lgE measurement
Number at percentage by
birth (%) levels of CS-lgE 2
<0.5 205 X p‘
value
Variable Total 1358 87.2 12.8
Gender Male 696 (48.7) 86.6 13.4
Female 661 (51 .2) 87.8 12.2 0.54
Missing 1 (0.1) 100 0.0
Birth weight <15009 27 (2.0) 88.9 11.1
1500-25009 44 (3.3) 87.2 12.8 0.76
2500-40009 1 124 (82.8) 90.9 9.1
40009< 163 (12.0) 85.3 14.7
Neonatal Yes 188 (13.8) 85.6 14.4 0.76
complications No 1 170 (86.2) 87.4 12.6
Having a pet at Yes 779 (56.6) 87.0 13.0 0.81
recruitment No 589 (43.4) 87.4 12.6
Season of birth Spring 315 (56.3) 87.2 12.8
Summer 305 (56.3) 87.9 12.1 0.1 3
Fall 296 (56.3) 83.1 16.9
Winter 442 (56.3) 88.5 11.5
Age of mother Up to 23 389 (28.6) 86.1 13.9
at delivery 24-33 years 801 (59.0) 87.4 12.6 0 54
34 and older 145 (10.7) 89.7 10.3 '
Missing 23 (1.7) 82.6 17.4
21:32:: :25 332 (24.6) 87.2 12.8
during Missing 1012 (74.5) 87.1 12.9 0.99
12 (0.9) 91.7 8.3
pregnancy
Maternal history Yes 438 (32.3) 84.0 16.0
of atopy No 905 (66.6) 88.6 11.4 0.02
Missing 15 (1 .1) 93.3 6.7
Paternal history Yes 352 (25.9) 85.0 15.0
of atopy No 984 (72.4) 88.2 11.8 0.11
_ Missing 23 (1.7) 78.3 21.7

 

CS-lgE=cord serum immunoglobulin

77

Table 3.2.2 Number of children who had a measurement of cord serum lgE
level at birth and at different analyses, by cord serum lgE levels

 

 

 

. 1-2
Age Birth years 4 years 10 years
Outcome of interest — Asthma Asthma SPT Asthma SPT
T°ta'"”mbe’°f 1358 1194 1068 862 1191 900
sample

1 1 84 1 044 928 748 1 040 789

N”mber(%)by (0'5 (87.2) (87.4) (86.9) (88.8) (87.3) (87.7)

different levels of

Cordserum|gE 205 174 150 140 114 151 111
' (12.8) (12.4) (13.1) (13.2) (12.7) (12.3)

 

SPT=skin prick test

78

0000 x25 :20 s.

 

 

 

 

 

 

 

.. 82 NE -- mom NS -- $0 new -- m: 3; A5 .98
84 mm? mm mm. a: E E4 mm mv 9: mm 5 omdm
34 En mm woe won an «we mom Nor ova mum 5 Ammdmv

mod
N: E3 mm 3... can mm mm; mum vmr Re 5N on :E: 5500005 A
S - + 5 - + m... - + y... - + 3va mm. 28
2. 0mm #0 mE£m< v 0mm 00 0E£m< or 0mm .0 .50 v 0mm 00 .Ew

 

mm. E200 Eco go 000. E0L0t5 5.. mE£mm can 9. .Eg 5:350:00 390:0 0o 8.: 00:9 085.05.. m.~.m 030...

79

.8. 00.5 50.0"»:0 .89 8.84.0
”.0295 00:00::0005 .mo .vnr E05500. :0 :00 0 0:30: 0:0 0:0:00._..E00 .0:0:00: 0:0 £0.02, :0... .5.... :0 :00000
$090 .0590: 0:0 .0E0:0E 502.00 :0 .050E :0 000 $8030... 0550 0:0.0E0 .0::0:0E 50:00 :0: m:...0.::00 ._.

 

 

 

 

8.0.8.8 2:880 $33.80 .85: $0.: 52.88.86
$0.18....m3 8.0.0-3903 32.0.8.0 8:00.883 0.5582
8.83.880 8.0-2.530 2:5...E0 8838.0 8:
8.0-8.83: 80.30.98: 3:580 .8398: 08 .80 28.2
5.0-8.580 .8018: EN 5:00.848 60:85:: 8.5.. 08.0
8.0-8.88: 0.8-8.803 33.8.8808 8:238? 2.5.08 88:
00.0-9.5 0: .80808: 3:04.380 850088 2:.
8.0-8.88: 5.0-8.903 3:30.80 .mvtosgtu 52.8885
8.98.520 8.3.5300 €53.80: 80:83.83 0.5582
5.0.8.380 8.0-3.980 €30.80 .85:§.~ 8:
80.98.52,.” 8.0-8.380 $2.008: 83.8.8.4 N8 .00
8:18:80 .0318: 8.0 3:58.: 84:00.80.» 8:8. 890
8.0-8.351 80:098.: $205.80: 64:808.: 9.5.08 88...
.80-?580 .808080 328.388 84:85.88: .5. 0.8....

._0 38. mo ..0 $8. :0 eyflfiymw ...%.m..m\v° .2. E0 8.:
._.00:0:._0< 0003005.. 0.0E00 .08...

0.00. mm. E200 0:00

 

.0.0>.0:< :0_000.m0m 0:050: 005. mm.-wo 0005.0 5.3 :0:00:.0:00 :0 0.0: 0.: :0: .0 0:0 0mO v.~.m 0.00..

80

:00. 0.0.... :.v.0u...n.0 ”0:0. 0000um0 ..0>.0::. 00:00::00u.0 .mo .vQ.
.050 0.: :0 m:.00.:. 0:0 000 0:0 :0 ._.n.w 0>:0m0: 00 :02. 00 or 0:0 0 000 .0 :0:0>.0000 m:.00.:. 00020:. a
€955.00. .0 .00 0 0:30.. 0:0 0:0:00.....:00 .0:0:00: 0:0 4:90.... 5.... .5.... :0 :00000
$000 .0500: 0:0 .0590... .>.0>_.00 .0 .059: .0 000 «6:030... 05.00 90.080 .0:.0.0:. ..00:0m .0. m:...0..:00 ._.

 

 

 

 

 

 

 

.598... K: .8080. 8: 20.0. 8.: £8.00. .8. 03 mefimwfis
28.03.: 8.0 .808: 8.0 .828: 08.0 .88. .80: 8v E0 02.082 .:
5.73.0.8... .8:-$.e8.o .800: .88 .83: 08.5 80 E0 02.000
..8.~-0o.:8: ....0.~-B.:8: 30:8. 08.0: 63:8: 08.: 3: .3
80.0-3... .0 .o 30.0-8.0: 3.0 .00.: .80 .85: 08.0 80 8.89.8 E0 02
8.73.0.8... .3753. 8 .o .080: 08.: .05: $5.: 80 E0 @5082 v
60.73880 .3780. E... .850. 08.8 .80: .88 0: E0 02.000
8.78.0.8... 8.78.88... 83:: 08.0: .083: 0.0.0: 8.: 30.
8.78.98... 5.73.88... 60:0: :8 .0328: $0.: 3: .2 0-:
c 0 c 0
:0 8mm: «.0 :0 $00: «.0 d2....v30. Dom. Afly“; OOH. 0..“7....v00 0:0..0 0m<
#000300. 0003005 0.0>0. mm. £3.00 0.00 .06...
.0_0>_0:< :0.000.o0m 0:050: 0.0>0. m.m.-00 00.0>0.0 5.2. 08500 .0 0.0.. 0.: .0: .0 0:0 0m0 0N...” 0.00...

81

.0.0>.0::. 00:00::00 #00 2000.00. 0.0: .00:.0> 0:: 0:0 00:0. 0000 0:: 30:0 0:00 0: ._. .m. :..::0.00::EE. E200
0.00 :0 0.0>0. 00:0>0.0 .0: 0000 ::0.0::_0 :0 0E::00 0:0 C00: :00: 0.0:: :_0.0 :0 $0: 00:0. 0000 00:03:00. _..N.m 0.00.“.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.00> o: v 0-:
o
F
V
_ . - u\ N 0.
— \\\ Nu?
_ \\
. \\!\ m m.
4. .50 m.
. v 0
_ m...
u m m.
_ I--- 2.308.303.0802 s
u ............. o:.0:.000:0..h.0 9208': 0
_ "9.0.2.52
: n
_
_
m

82

m. :..::0.00:::..:_ E200

0.00 :50. 0.0 m. 00:0>0.0 :0 000.520.: 0:0 0000 ::0.0::.0 :_ 0000:00_0 00:08:00 :0 0000.0 .0: 0.80:5 0.0.0 0.30.“.

 

 

 

 

mEA me: A 000 A
E \\ E \\
.0 8.
mm: 4| 00 8 All: on Illum—
\
E \
5
mm mm
0.00.. 0.. 0.00.. v 0.00.. N-..

 

 

 

00.: 00.: 0650».

223

08500 .0000. 02
5.5

 

0MV

83

3.3. The combined effect of maternal smoking during pregnancy

and IL13 gene polymorphism on childhood wheeze phenotypes

84

Abstract

Tobacco smoke and genetic susceptibility are risk factors for asthma and
wheezing. The aim of this study was to investigate whether there is a combined
effect of tobacco smoke and the interleukin 13 gene (IL 13) polymorphisms on
wheezing. In the Isle of Wight birth cohort (UK, 1989-1999), five single nucleotide
polymorphisms were genotyped. Wheeze phenotypes were early-onset
persistent (onset in the first 4 years, present at age 10); late-onset persistent
(onset after age 4 years, present at age10); eady transient (only in the first 4
years); and non-wheezers. Logistic regression analyses included 787 children.
Controlling for confounders, maternal smoking during pregnancy was associated
with early-onset persistent wheeze (odds ratio (OR) = 3.05; p < 0.001 ); the
common IL13 haplotype pair was not (OR = 1.01; p = 0.95). However, the
association of smoking was stronger in children with the common IL13 haplotype
pair (OR = 5.54) compared to those without it (OR = 1.67; p for interaction =
0.026). This is the first report that shows a combined effect of in utero exposure
to smoking and IL13 on wheezing in childhood. The results emphasize that
genetic association studies need to take exposures into account, since they may

explain contradictory findings.

85

It has been suggested that the increased prevalence of asthma-related
phenotypes over the last three decades is due to exposure to environmental
factors (95). Among such exposures, cigarette smoking is regarded as an
important risk factor for asthma (22). In particular, an effect of maternal smoking
during pregnancy on the development of asthma in offspring has been proposed
(96—102). A number of studies in children have linked parental smoking with
markers of atopy such as serum immunoglobulin E (lgE) levels (103, 104).
Interleukin 13 (IL-13) is an important cytokine involved in the lgE pathway and
pathogenesis of asthma (35, 105, 106). Noakes and colleagues noted that
exposure to tobacco smoke during pregnancy was associated with a significantly
higher neonatal lL-13 response to both house dust mite and ovalbumin,
indicating that a prenatal exposure to tobacco may have immunological effects
(105). These observations suggest that part of the effect of exposure to tobacco
on asthma might be through cytokines.

Researchers have speculated that individuals may vary in genetic
susceptibility to cigarette smoke exposure (104, 107) and a few studies have
examined gene-smoking interactions for asthma (30, 40, 41 ). Wang et al.
observed that an interaction between maternal smoking and the 02 adrenergic
receptor gene was associated with asthma in offspring (41). Kabesch et al.
reported an effect for the interaction between smoking and glutathione S
transferase deficiency on asthma (40). In spite of the importance of lL-13 in
asthma (35, 105, 106), some studies failed to show an association between the

interleukin 13 gene (IL13) polymorphism and asthma phenotypes (108, 109) and

86

there is no information as to whether the adverse effect of maternal smoking on
asthma may result from an interaction between exposure to tobacco smoke and
cytokine regulating genes. To the best of our knowledge, no study has
investigated whether there is an interaction between tobacco smoke exposure
and IL13 polymorohism on wheezing. Because of the high prevalence (25-30
percent) of smoking among women of reproductive age (110) there is a need to
investigate its possible role in the development of asthma and its related
phenotypes. In this work the authors test the hypothesis that an interaction
between exposure to tobacco smoke and IL13 gene polymorphism affects
longitudinal assessments of wheeze phenotypes during the first decade of life in
the Isle of Wight birth cohort.
POPULATION AND METHODS
Study population

The local Research Ethics Committee approved the study of children born
and enrolled (n=1,456) between January 1, 1989, and February 28, 1990, on the
Isle of Wight, United Kingdom. lnforrned consent was obtained and children were
followed up at the ages of one (n = 1,167; 80.2 percent), two (n = 1,174; 80.6
percent), four (n = 1,218; 83.7 percent) and 10 years (n = 1,373; 94.3 percent).
The island is close to the British mainland, semi-rural, with no heavy industry.
The population is 99 percent Caucasian.
Family history, birth data, and follow-ups

At birth, weight was obtained from hospital records and parents were

asked about their age and exposure during pregnancy (maternal smoking and

87

household pets). At subsequent follow-ups, a questionnaire was administered
seeking information on exposure to tobacco smoke, breast feeding, and
symptoms that were suggestive of allergic diseases in the cohort children.
Exposure to tobacco smoke

Information on tobacco smoking by mother (during pregnancy and later),
father, or any other individual inside the home was recorded at recruitment and
updated at each follow-up. Exposures to environmental tobacco smoke (ETS) in
the household and maternal smoking during pregnancy were combined and
classified into three groups. When mothers did not smoke during pregnancy and
there was no exposure to household ETS in children up to the age of 10, children
were categorized as “ETS-0". When mothers did not smoke during pregnancy,
but household members smoked within the home at some point up to the child’s
age of 10 years, the exposure status was categorized as “ETS-1”. When mothers
smoked during pregnancy and the children were also exposed to household ETS
at some point up to the age of 10, the exposure was categorized as “ETS-2”.
None of the children had mothers who smoked during pregnancy but no
exposure to household tobacco smoke after birth.
Defining of wheeze and wheeze phenotypes

At each of the four follow-ups at ages one, two, four, and 10 years,
“current wheeze” was recorded if wheezing occurred on at least one occasion in
the previous 12 months. Parents were also asked whether the child had any
wheeze episode between ages four and 10. Children were categorized into

wheezing phenotypes using the information on “current wheeze” at each follow-

88

up and also whether they had wheeze episode(s) between ages 4 and 10 years.
Non-wheezers never wheezed in the first decade of life, eady-onset persistent
wheezers had wheezed in the first 4 years of life and still wheezed at age 10,
while late-onset persistent wheezers had wheezing onset after 4 years of age
and still wheezed at age 10. Early transient wheezers began wheezing in the first
4 years but stopped at least 12 months before age 10. Details on wheeze
phenotypes from 1 to 10 years of this cohort have been reported previously (15).
DNA isolation and IL13 genotyping

Anticoagulated whole blood samples were obtained at the 10-year
interview and stored frozen (n=921). Genomic DNA was isolated from these
samples using QIAamp DNA Blood Kits (Qiagen, Valencia, CA) or the ABI
PRISM” 6100 Nucleic Acid PrepStation (Applied Biosystems, Foster City, CA).
Polymorphisms in the IL 13 gene were examined using the SNPper (120) and
Applied Biosystems (121) databases. Two methods were used for genotyping: 1)
fluorogenic 5’ nuclease chemistry PCR using Assays on Demands kits cycled on
a 7900HT Sequence Detection System (Applied Biosystems, Foster City, CA),
and 2) biotin-streptavidin-based pyrosequencing performed on PSQ-96
instrumentation (Biotage AB, Uppsala, Sweden).

Five SNPs from the IL13 gene were used in this study, rs1800925 in the 5’
promoter region, rs2066960 in intron 1, rs1295686 in intron 3, r320541 in exon 4
and rs1295685 in the 3’ untranslated region (3’ UTR) of exon 4. Because IL13 is
a small gene (2.9 kb), only a few SNPs were needed for a reasonable

assessment of genetic associations. Polymorphism selection was based on SNP

89

location, minor allele frequency, and function. Selected SNPs had minor allele
frequencies 2 19% and were distributed throughout the IL13 gene. The rs20541
is a coding variant with the common allele (G) coding for arginine and the minor
allele (A) encoding glutamine at amino acid 144. This SNP was in strong linkage
disequilibrium (LD) with rs1295685 and rs1295686. The former (rs1295685) was
selected as a representative SNP of this block due to the availability of a
commercial assay.
Statistical analysis

Each SNP was tested for Hardy-Weinberg equilibrium using Haploview 3.2
software (124). Estimates of LD between SNPs were calculated using 0' and r2
(116). For SNPs that were in strong LD, the authors took one SNP as the
representative marker for the block. PHASE 2.0.2 was employed to build the
most likely pair of haplotypes (diplotypes) and their probability for each child (117,
125), using representative SNPs. In the analysis of individual SNPs and
haplotype pairs, genotypes with low frequencies were combined (126, 127).

Using SAS/STAT® version 9.1, statistical analysis was performed on the
data only from children who had complete information on IL13 genotypes,
exposure to tobacco smoke, and wheeze phenotypes. Chi-square tests were
used to compare the sample used in the analysis with that which was followed up
at age 10. Logistic regression analysis and Wald tests were conducted to
investigate the statistical interaction between ETS and IL13 haplotype pairs on
wheeze up to age 10.

Statistical significance was defined as a p value of s 0.05. As the four

90

wheeze phenotypes were not independent and we focused on only one
haplotype pair, no adjustment for multiple testing in the haplotype pair analysis
was performed (165).

For confounders, the authors considered covariates that have been
suggested to be important in studying wheeze phenotypes in a prior report of the
Isle of Wight cohort. (166) Potential confounders (table 3.3.1) were: gender, low
birth weight (< 2,500 grams versus 2 2,500 grams), breast feeding (< three
months and 2 three months), household cat present during pregnancy (yes, no),
and household dog present during pregnancy (yes, no).

RESULTS

Data on IL13 genotypes, exposure to tobacco smoke, and wheeze
phenotypes were available for 787 children. The percentages of children who
wheezed and those who had ETS-0 were higher in the sample used in the
analyses compared to all those who were followed up to age 10 (p values of
0.006 and 0.003, respectively, table 3.3.1).

More than half of the children were exposed to tobacco smoke up to age
10 (20.2 percent during and after pregnancy; 31.5 percent after pregnancy, table
3.3.1). In the sample used for the analysis, the odds ratio for ETS-2 (reference
group: ETS-0) for early-onset persistent wheeze was 3.05, 95 percent confidence
interval: 1.67, 5.57. To investigate whether there was a selection bias due to
tobacco smoke exposure and wheeze, the authors calculated the odds ratio for
their association in all children with a follow-up at age 10 (2.93, 95 percent

confidence interval: 1.76, 4.89).

91

For IL13 genotypes, all five SNPs were in Hardy-Weinberg equilibrium (p
= 0.82 for rs1800925, p = 0.99 for r82066960, p = 0.58 for rs1295686, p = 0.09
for r320541, and p = 0.36 for rs1295685). Three of the SNPs (rs1295686,
rs20541, and rs1295685) were in strong LD (figure 3.3.1). Therefore the authors
used only one SNP of this block for haplotype analysis. Thus haplotypes were
inferred using rs1800925, rs2066960 and rs1295685. Due to the limited number
of children homozygous for minor alleles at all SNPs (less than 5 percent, table
3.3.2), minor allele homozygous and heterozygous genotypes were grouped
together. Among haplotype pairs, CCG/CCG had the highest frequency (0.48,
table 3.3.3). The estimated probability for CCG/CCG as the best pair was 1.0 in
99 percent of children and was never less than 0.89 (results not shown). All other
haplotype pairs had frequencies between 0.001-0.151. In addition, there were
probabilities as low as 0.5 that were identified as the best pair for haplotype pairs
other than CCG/CCG. Thus, haplotype pairs other than 006/ CCG were
combined (126, 127).

Higher percentages of early-onset persistent wheeze were observed when
children had exposure to maternal smoking during pregnancy (ETS-2) and the
major genotype (major allele homozygous genotype of the SNPs or CCG/CCG of
haplotype pairs, figure 3.3.2). In a logistic regression model, controlling for the
confounders, ETS-2 was associated with earty-onset persistent wheeze (odds
ratio = 3.05, 95 percent confidence interval: 1.67, 5.56), but the CCG/CCG
haplotype pair was not (odds ratio = 1.01, 95 percent confidence interval: 0.64,

1.61 ). References for ETS-2 and CCG/CCG haplotype pairs were ETS-0 and

92

other haplotype pairs, respectively. When the interaction term between ETS-2
and CCG/CCG was added to the model, the term showed statistical significance
(odds ratio = 3.47; p < 0.026, table 3.3.4). To demonstrate the interaction
between ETS-2 and the CCG/CCG haplotype pair on earty-onset persistent
wheeze, the authors stratified the sample by the IL13 polymorphisms and
calculated the odds ratios for ETS-2. Children who had the CCG/CCG genotype
had an odds ratio = 5.54, 95 percent confidence interval: 2.28, 13.50 for ETS-2.
The odds ratio in the stratum that had haplotype pairs other than CCG/CCG was
1.65, 95 percent confidence interval: 0.70, 3.91. To determine the effect of minor
haplotype pairs (frequency= 0.082, table 3.3.3), the stratified analysis was
repeated by excluding this group and results did not change significantly (data
not shown).

Among confounders, the presence of a household cat during pregnancy
was significantly associated with a lower risk of early-onset persistent wheeze
(odds ratio = 0.51, 95 percent confidence interval: 0.30, 0.87), but not
significantly associated with other wheezing phenotypes (odds ratio = 0.98 and
0.90 for early transient and late-onset persistent wheeze, respectively). Breast
feeding for at least three months was protective for early transient wheeze (odds
ratio = 0.68, 95 percent confidence interval: 0.46, 0.99) and for early-onset
persistent wheeze (odds ratio = 0.67, 95 percent confidence interval: 0.41, 1.09),
but not for late-onset persistent wheeze (odds ratio = 1.04, 95 percent confidence
interval: 0.61, 1.78). Boys were at a higher relative risk of being among wheeze

phenotypes, especially, of having early-onset persistent wheeze (odds ratio =

93

1.99, 95 percent confidence interval: 1.24, 3.21 ). The odds ratios for early
transient wheeze and late-onset persistent wheeze were 1.56, 95 percent
confidence interval: 1.08, 2.26 and 1.31, 95 percent confidence interval: 0.77,
2.24, respectively.

DISCUSSION

This study investigated the combined effect of exposure to tobacco smoke
and haplotype pairs of the IL 13 gene on childhood wheeze phenotypes in the first
decade of life using the data from the Isle of Wight birth cohort. No independent
effect for the IL13 gene was detected, but the common variant of the IL13 gene
polymorphism was shown to increase the adverse effect of maternal smoking
during pregnancy (but not tobacco smoke exposure after pregnancy) on early-
onset persistent wheeze. It was previously shown in this cohort that persistent
wheezing with early childhood onset is associated with substantial morbidity in
the first decade of life in association with high levels of atopy, bronchial hyper-
responsiveness and impaired lung function at 10-years of age (15).

For this study, the information was available from a subset of children who
were followed up at age 10 and who agreed to provide blood for genotyping.
These children appeared to have more wheeze in comparison to all children who
were followed up at age 10. In addition, the percentage of mothers who never
smoked was higher in the group who provided a blood sample for genotyping.
Therefore, a selection bias may exist. As presented in the results section, the
odds ratios for the association between exposure to tobacco smoke during

pregnancy and early-onset persistent wheeze in those who were used in the

94

analysis and all children (3.05 and 2.93, respectively) are nearly identical. This
suggests that a selection bias for tobacco exposure pattern, if it exists, is
negligible. In addition and with regard to genotypes, presence of a selection bias
could result in a violation of the Hardy-Weinberg equilibrium. The genotypes of
the five SNPs used in the analysis were in Hardy-Weinberg equilibrium and their
allele frequencies were comparable to the equivalent SNPs in other Caucasian
populations (72, 73). Hence, a selection bias is unlikely.

The wheeze phenotypes were specified in an assessment that was
performed at age 10. A strength of this study is that the determination of wheeze
phenotype is based on longitudinal records from ages one to 10 years. To avoid
recall bias, as explained in more detail in a prior report of this cohort (15), the
analysis was restricted to children who were seen prospectively with information
at all study visits. Previously, the authors have shown that a prior diagnosis of
asthma was less likely to produce biased reports (119).

With regard to genetic analysis, information on individual SNPs was used
and the most likely pairs of haplotypes were estimated from genotype data. It
was possible to use information on each individual SNP to investigate gene-
environment interactions. The authors chose to perform tests of interaction for
haplotype pairs, as this approach has some advantages over individual SNP
analysis. The issue of “multiple testing” is avoided as all genetic markers were
compiled in haplotype pairs. Additionally, it has been suggested, specifically for
IL 13 (132), that haplotype analysis could confer more information than individual

marker analysis (132-134). Haplotype pair analysis may misclassify genotypes

95

when parents’ genetic information is not available (ambiguous phase). However,
the probability of having CCG/CCG, the major haplotype pair, was 1.00 in 430
out of 435 children with this genotype (the probability for the other 5 children was
0.89). Also, the consistent pattern observed for the combined effect of ETS-2 and
all markers (individual SNPs polymorphisms and haplotype pairs) on early-onset
persistent wheeze (figure 3.3.2) suggests that misclassification of the genotypes
was unlikely. The distribution of the data, with respect to minor haplotypes, did
not allow testing for their interactions with tobacco smoke exposure or specifying
a genetic model (additive, recessive or dominant). However, when children with a
minor haplotype at both loci (frequency = 0.082, table 3.3.3) were removed from
the analysis, the results did not change substantially.

Previous studies have shown an increased risk of asthma in children who
were exposed to tobacco during pregnancy (22, 101, 107). In a prior examination
of this cohort, a possible association between exposures during pregnancy and
early-onset persistent wheeze was suggested (166). Additionally, several studies
have suggested a gene-environment interaction for the effect of tobacco smoke
exposure and asthma phenotypes (30, 40, 167). The current study demonstrates
a scenario in which a gene modifies the effect of tobacco smoke exposure during
pregnancy but not after (the time of the exposure may be of critical importance in
gene-environment interaction studies). The finding of an interaction between IL13
and ETS in the present study suggests that negative reports for the effect of a
candidate gene (in this case, IL13 on asthma phenotypes (108, 109)), could be

explained by a failure to take into account environmental exposures. It is

96

therefore of utmost importance, for genetic association studies to describe
environmental exposures in the target population.

Some authors have advised against publishing results without replication
studies (168). In the present study there is no replication of the current findings in
another population. However, the authors believe that the replication requirement
is not a conditio sine qua non and should be considered in the context of external
validity. Other studies have reported similar combined effects of IL13
polymorphisms on asthma-related outcomes such as lgE, or smoking-related
effects on the lL-13 cytokine (105, 148). Hence, there is external validity for these
findings.

The mechanism behind the effect of tobacco exposure on childhood
asthma is not clear. Most speculation has centered on the suggestion that
tobacco smoke impairs lung development during infancy and early childhood and
therefore leads to asthma (96-102). It is well known that in addition to structural
defects, immune mechanisms and cytokines play prominent roles in asthma. IL-
13 is one of the cytokines that has a pivotal function in the pathogenesis of
asthma (6) and IL13 genetic polymorphisms have been shown to be related to
IgE and asthma (72, 73, 148). There is also evidence that tobacco smoke can
alter immune markers that are involved in allergic immune reactions. Liu et al.
reported a synergistic effect of smoking and IL13 promoter polymorphism on the
level of serum lgE (148), and a genome-wide analysis has suggested a gene-
environment interaction for the effect of tobacco exposure on asthma (30).

Noakes and colleagues demonstrated that cord blood cells produce significantly

97

higher levels of lL-13 in response to both house dust mite and ovalbumin when
newborns are exposed to tobacco smoke during pregnancy (105). This suggests
that prenatal exposure to tobacco elicits immunological effects. In our analysis,
the interaction between tobacco exposure and IL13 polymorphisms was present
in those who were exposed during and after pregnancy, but was not evident for
the group with the same polymorphism who were exposed only after pregnancy.
As there was no group of children that was exposed only during pregnancy, it is
not possible to distinguish definitively the effect of tobacco exposure before and
after pregnancy. However, considering both the relatively short time of pregnancy
and the large difference of risk between the two exposed groups, the authors
assume that exposure to tobacco during pregnancy has more influence on
asthma than tobacco exposure after pregnancy. There is evidence that
remarkable changes in maternal immune responses occur during pregnancy.
Also the fetal immune system is active as early as 10-22 weeks of gestation (62).
Hence, the fetal period may be a critical time for the development of immune-
related disorders later in life. It has been hypothesized that the dramatic increase
in the prevalence of asthma in the last three decades in developed countries is
because of environmental exposures. One possible explanation for the
differential effects of environmental exposures in populations could be gene-
environment interactions. There is a potential for genetic information to change
without a sequence variation in DNA. These changes happen through post
synthetic modifications of either DNA strands or histones and rest on

developmental and/or environmental insults and are known as epigenetic

98

phenomena (5,42). The results of the current study of the combined effect of in
utero exposure to tobacco and IL13 genotypes on wheeze could be explained by
such epigenetic changes (5,43).

In summary, in a sub-sample of the Isle of Wight cohort, the combined
effect of exposure to tobacco smoke during pregnancy and the common
haplotype pair of the IL13 gene resulted in an increased relative risk of early-
onset persistent wheeze. Tobacco smoke exposure later in childhood was less
important. The IL13 gene did not pose a risk in its own right. These results
demonstrate that the association between exposures to environmental risk
factors like tobacco smoke can be modified by gene polymorphisms. Given that
there are various patterns and prevalences of exposure to tobacco smoke in
different populations, this study suggests that negative reports of genetic
association studies, for example Leung et al., 2001 (108) and Hakonarson et al.,
2001 (109) for IL13 and asthma, may be due to differences in environmental
exposures. The authors propose that the next step in the investigation of the
interaction between IL13 and exposure to tobacco smoke is to examine the

expression of IL 13.

99

Table 3.3.1. Comparison of children with a follow-up at age 10 and the subset
used in the analysis

 

 

2321382; Tl‘fl’fiéiyéliegtl‘ .2 ,,
Variable Total 1 .373 787
Wheeze Eaflxflnset 139 (10.1) 105 (13.3)
Phenotype fersrstent“
819:00891 81 (5.9) 70 (8.9)
persrstent’r 0005
Early transient-t 259 (18.9) 188 (23.9)
Non-wheezer§ 617 (44.9) 424 (53.9)
Missing 277 (20.2) 0 (0.0)
Birth weight 22500 grams 1287 (93.7) 733 (93.1)
<2500 grams 49 (3.6) 30 (3.8) 0.553
Missing 37 (2.7) 24 (3.1)
Exposure to ETS-0 617 (44.9) 380 (48.3)
tobacco ETS-1 430 (31.3) 248 (31.5)
0.003
smoke ETS-2 319 (23.2) 159 (20.2)
Missing 7 (0.5) 0 (0)
Gender Girl 676 (49.2) 388 (49.3)
Boy 697 (50.8) 399 (50.7) 0'95“
Breast fed 23 months 564 (41.1) 360 (45.7)
<3 months 667 (48.6) 394 (50.1) 0.088
Missing 142 (10.3) 33 (4.2)
Household Yes 453 (33.0) 269 (34.2)
cat during No 912 (66.4) 517 (65.7) 0.343
pregnancy Missing 8 (0.6) 1 (0.1)
Household Yes 395 (28.8) 239 (30.4)
dog during No 970 (70.6) 547 (69.5) 0.163
pregnancy Missing 8 (0.6) 1 (0.1)

 

* Onset of wheeze in the first 4 years, still present at age 10.
1' Onset of wheeze after age 4 years, still present at age10.

:j: Wheeze only in the first 4 years.

§ No wheeze up to age 10.
ETS-0, mothers did not smoke during pregnancy and children not exposed to

household ETS; ETS-1, mothers did not smoke during pregnancy, but children
were exposed to household ETS; ETS-2, mothers smoked during pregnancy and
children were exposed to household ETS

100

Table 3.3.2. Genotypes for IL13 single nucleotide polymorphisms (SNPs)

 

 

SNP Position (bp) Location Genotype Frequency (%)
rs1800925 132,020,708 Promoter CC 577 (63.6)
CT 295 (32.5)
TT 35 (3.9)
Total 907 (100.0)
rs2066960 132,022,334 Intron 1 CC 729 (81.5)
AC 157 (17.6)
AA 8 (0.9)
Total 894 (100.0)
rs1295686 132,023,742 Intron 3 CC 483 (64.6)
CT 240 (32.1)
TT 25 (3.3)
Total 748 (100.0)
r820541 132,023,863 Exon 4 G6 73 (53.7)
AG 59 (43.4)
AA 4 (2.9)
Total 136 (100.0)
rs1295685 132,024,344 Exon 4 GG 584 (64.5)
AG 280 (30.9)
AA 41 (4.5)
Total 905 (100.0)

 

Gene annotation is based on SNPper (120).

IL13: interleukin 13 gene; n, number of samples used in genetic studies; bp,

basepairs.

101

Table 3.3.3. Frequency of haplotypes and haplotype pairs for three single
nucleotide polymorphisms (rs1800925, rs2066960, rs1295685) of the
interleukin 13 gene, inferred from genotype data of 907 children

 

 

Frequency (SE)

Haplotype CCG 0.683 (0.004)

CCA 0.045 (0.002)

CTG 0.067 (0.003)

CTA 0.108 (0.003)

ACG 0.041 (0.003)

ACA 0.029 (0.002)

ATG 0.008 (0.002)

ATA 0.017 (0.003)
Haplotype pair(8) CCG/CCG 0.480
CCG/minor haplotypes* 0.438

minor haplotypes/minor haplotypesl' 0.082

 

PHASE 2.0.2. was used for this analysis (117, 125).

* This group consists of, with a frequency of:

CCG/ACA, 0.036; CCG/ACG, 0.055; CCG/ATA, 0.037; CCG/ATG, 0.019;
CCG/CCA, 0.054; CCG/CTA, 0.151; CCG/CTG, 0.086.

T This group consists of, with a frequency of:

ACA/ATA, 0.001; ACG/ACA, 0.003; ACG/ACG, 0.001; ACG/ATA, 0.002;
ATG/ATA, 0.001; CCA/ACA, 0.004; CCA/CCA, 0.002; CCA/CTA, 0.014;
CTA/ACA, 0.014; CTA/ATA, 0.002; CTA/CTA, 0.007; CTG/ATA, 0.004;
CTG/CTA, 0.020; CTG/CTG, 0.004.

102

Table 3.3.4. Logistic regression estimates* (odds ratio, p value) for the effect of
smoke exposure and the major haplotype pair (CCG/CCG)T of the interleukin 13

gene on wheeze phenotypes

 

Wheeze phenotype: Odds ratio§ p value
Early—onset persistent, n=105
ETS-1 1.64 0.186
ETS-2 1.57 0.294
CCG/CCG 0.58 0.168
ETS-1 x CCGICCG§ 1.58 0.370
ETS-2 x CCGICCG§ 3.47 0.026
Late-onset persistent, n=70
ETS-1 0.78 0.559
ETS-2 0.39 0.159
CCG/CCG 0.90 0.766
ETS-1 x CCGICCG§ 0.67 0.929
ETS-2 x CCGICCG§ 0.65 0.499
Early transient, n=188
ETS-1 1 .42 0.248
ETS-2 1.84 0.067
CCG/CCG 0.95 0.854
ETS-1 x CCGICCG§ 1.69 0.609
ETS-2 x CCGICCG§ 1.33 0.462

 

* Adjusted for gender, birth weight, duration of breast feeding, having a cat
during pregnancy, and having a dog during pregnancy.

1 The most common haplotype pair, inferred from genotype data of three single
nucleotide polymorphisms of interleukin 13 gene, rs1800925, r52066960 and
rs1295685, using PHASE 2.0.2 software (117, 125).

1: The reference was the group who never report wheeze up to age 10 (n=424);
x Denotes an interaction.

§ Odds ratios for interactions were calculated by exponentiating the sum of the
main effect terms and the interaction term; for example for ETS-2 x CCG/CCG
in early-onset persistent phenotype group, odds ratio = exponential (0.56 — 0.45
+ 1.34) = 3.47

ETS-0, mothers did not smoke during pregnancy and children not exposed to
household ETS; ETS-1, mothers did not smoke during pregnancy, but children
were exposed to household ETS; ETS-2, mothers smoked during pregnancy
and children were exposed to household ETS

103

"'5

rs1800925
rs2066960

54

‘\

rs1295685

 

Figure 3.3.1. Using the Haploview program (124),
three single nucleotide polymorphisms of the
interleukin 13 gene appeared to be in a block.

 

Pair of SNPs

’2

D’ (95% Cl)

 

rs1800925 - r52066960:
rs1800925 — rs1295686:

rs1800925 — r820541:

rs1800925 — rs1295685:
r82066960 — rs1295686:

r52066960 — r520541:

r52066960 — rs1295685:

rs1295686 — r820541:

r81295686 — rs1295685:
r320541 — rs1295685:

0.01
0.31
0.28
0.28
0.06
0.07
0.06
0.78
0.90
0.85

0.14 (0.04, 0.24)
0.56 (0.50, 0.62)
0.63 (0.46, 0.75)
0.54 (0.48, 0.59)
0.37 (0.27, 0.47)
0.37 (0.04, 0.71)
0.38 (0.28, 0.47)
1.00 (0.83, 1.00)
0.96 (0.93, 0.98)
1.00 (0.94, 1.00)

 

D’ and r2 are pair wise linkage disequilibrium
determinants (116); CI, confidence interval.

104

50 Cl Others [Major genotype*

45
4O
35
30
25
20
15
10

 

 

Early-onset persistent, P = 105I 529

 

50
45
4O
35
30
25
20 — "

%

 

 

70/ 494

 

 

 

 

 

 

 

 

 

10-
5.1

 

 

 

 

 

 

 

 

Late-onset persistent, P
a.“
3

Early transient, P = 188/ 612

 

9 7 9 9 ‘7 9 9' ‘7 9 9' ‘7 9
‘12 ‘12 9 ‘12 ‘12 E ‘1’2 I’.’ I’3 1’3 ‘12 1‘2
Ill 01 DJ [LI LU In DJ m LU 01 w I.l.l
Marker1': rs1800925 rs2066960 rs1295685 Haplotype pairs

Figure 3.3.2. Proportions of children (P) with a specific wheeze phenotype:
number of children with each wheeze phenotype divided by the sum of children
with the respective phenotype plus 424 non-wheezers.

105

Figure 3.3.2. (cont’d) ETS-0, mothers did not smoke during pregnancy and
children were not exposed to household ETS; ETS-1, mothers did not smoke
during pregnancy, but children were exposed to household ETS; ETS-2, mothers
smoked during pregnancy and children were exposed to household ETS.

SNP, single nucleotide polymorphism; * For each marker, “major genotype” was
the genotype with the highest frequency and “others” was a combination of all
other genotypes; 1' rs1800925, rs2066960, and rs1295685 were three SNPs
within the interleukin13 gene, “haplotype pairs" were constructed based on these
SNPS, using PHASE 2.0.2 software (117, 125).

106

DISCUSSION

107

Using data from the Isle of Wight birth cohort study, I demonstrated first, a
modifying effect of maternal smoking on the association between the common
variant of the IL13 gene and levels of cord serum immunoglobulin E (CS-lgE);
second, an association between CS-IgE and childhood asthma; third, a
combined effect of the common IL13 variant and maternal smoking during
pregnancy on the early onset persistent wheeze phenotype.

The motivation to conduct this research was to investigate whether
exposures during pregnancy can have a Iong-terrn effect on health outcomes, the
prenatal programming concept. The conceptual model for the long-lasting effect
of exposures during pregnancy on diseases was exemplified by using tobacco
smoke during pregnancy and IL13 polymorphism as risk factors, CS-lgE as a
marker of susceptibility for asthma, and asthma phenotypes as outcomes.

The mechanism for prenatal programming explored in the first and third
papers was an interaction between IL 13 gene polymorphism, a predisposing
factor, and exposure to tobacco smoke during pregnancy. In other words, the
combined effect of the common IL13 variant and maternal smoking during
pregnancy was observed for both CS-IgE and persistent wheezing. The
association between CS-lgE and asthma, the second manuscript, showed that
the endpoints (CS-lgE and persistent wheeze) of the above-mentioned gene-
environment interaction are related.

The three pieces of this work suggest that in utero exposure to tobacco
smoke may alter the expression of the IL13 gene at birth, a change that may last

as late as age 10. The interaction of IL13 polymorphism was not observed in

108

I -

 

those who were exposed to tobacco smoke after birth unless they had exposure
during pregnancy. This suggests that the in utero environment has a critical
effect on the pathogenesis of asthma and that the prenatal period is of utmost
importance.

To the best of my knowledge there is no investigation on the interactions
between genes and prenatal exposures on childhood asthma. Barker was one of
the first authors to raise the idea of prenatal priming of adulthood diseases. (169) 13'
His observations and their interpretations involve the effect of fetal malnutrition
on the development of organs; they provide no explanation for effects at the
molecular level. (169-171)

Regarding the puzzle of an effect of prenatal exposures on the
development of asthma, several pieces of evidence exist, but they have not been
explored together in one study. There is certain evidence that exposures during
pregnancy can change newborns’ immune markers that are involved in the
biologic pathways that lead to asthma. (58, 66, 67, 172) Recently a large meta-
analysis study has shown that the decrease in lung function parameters in
children aged 6 to 12 years is greater when they had in utero exposure to
maternal smoking compared to after birth exposure to tobacco smoke. (31)
However, there is insufficient information on whether and how these changes
may last until late childhood. One obstacle is that usually those children whose
mother smoked during pregnancy also have exposure to tobacco smoke after
pregnancy. For example, in the Isle of Wight cohort, there was no group of

children who were exposed to maternal smoking during pregnancy but not to

109

 

tobacco smoke after birth. Therefore the evaluation for the effect of maternal
smoking during pregnancy was only indirect by statistical modeling and
controlling for the exposure to tobacco smoke after birth. Future studies may
overcome this limitation by approaching a group of children that had exposure
during pregnancy but not after, for example exposed children who were adopted
by non-smoking parents shortly after birth.

With regard to the interaction between the IL13 gene and environmental

factors on serum lgE, the only report demonstrated a combined effect of one

SNP in the IL13 gene (rs20541) and exposure to tobacco smoke. (148) It has
been suggested that smoking can change gene expression through chemical
alteration, for example acetylation of chromatin (149, 173-175). There is no study
that determines whether exposure to tobacco smoke can alter IL13 expression,
and if so, what mechanism is involved. Therefore, future studies are warranted to
explore, in the first step, the IL13 expression under different conditions of
exposure to tobacco smoke; and, in the second step, the specific mechanism
behind the effect of smoking on IL13 gene expression.

The aim of this dissertation was to coalesce pieces of the prenatal
programming concept of asthma. The longitudinal nature of the Isle of Wight
study offered the opportunity to investigate the interaction between the IL13 gene
and maternal smoking during pregnancy on an asthma susceptibility factor (cord
serum lgE) at birth and to track this gene-environment interaction effect up to 10
years of age, for which a similar interaction was observed for persistent wheeze.

Therefore, this investigation provides intemal validity for the studied gene-

110

 

environment interaction under consideration as a mechanism for prenatal
programming.

In addition to the general advantages of a cohort study, such as the minimal
amount of recall bias, longitudinal studies possess other strengths that are
specific to the current study. The longitudinal nature of the data not only enabled
me to connect the three pieces of the study mentioned above, but also provided
the opportunity to obtain a longitudinal evaluation of exposure to tobacco smoke
and wheezing phenotypes in the first decade of life. The latter benefit is an
important aspect for asthma diagnosis because childhood asthma is a dynamic
disease with considerable heterogeneity and can best be assessed through a
cohort study.

Another strength of this study was a thorough assessment of the IL13 gene
polymorphism. The IL13 gene is a relatively small gene and the five SNPs
provided adequate representation of the IL 13 gene variability. The haplotype pair
analysis approach empowered the analyses as the number of comparisons
diminished.

The general limitation for gene—environment interaction studies is sample
size. As the frequency for individual variants of a gene decreases or the number
of these variants increases, the available sample size with those variants gets
smaller. The sample size problem becomes more prominent when the research
focuses on a specific time window for exposure, such as the prenatal period. In
order to overcome the problems related to sample size, most of the genetic

studies used a case control design rather than a cohort. Despite the advantage

111

 

 

of case control studies, larger sample size, this design usually suffers from recall
bias. For example, in the context of the current study, for persistent wheezing
and tobacco smoke exposure at different ages, it is unlikely to obtain valid
information, not affected by recall bias, within a case control design. With regard
to the IL13 polymorphism in the current study, it was not possible to test the
interaction effect for some of the variants in the gene. However, when the
variants with small sample size were removed from the analyses, the change in
the results was not considerable.

Epidemiology is the art and science of providing valid clues to the cause of
a health-related problem. This dissertation met this definition by investigating an
under-explored mechanism for the pathogenesis of childhood asthma. Although
there is evidence for the adverse effect of smoking on childhood asthma, the
mechanism behind this effect is unclear. The findings of the current work present
a scenario in which a prenatal exposure (tobacco smoke) may induce a long-
Iasting change in a susceptibility factor (IL13) for asthma, the most common
chronic illness of children. If this work is corroborated, it may then be found that
the adverse effect of maternal smoking during pregnancy may last until
adulthood and even transfer to offspring of the exposed individuals via epigenetic
mechanisms. (44,45)

It is generally believed that the recent increase of asthma and allergy is
related to environmental pollutants. However, the idea of the in utero
programming of asthma has not been explored either for exposure to tobacco or

for any other environmental exposure. Some of these pollutants, for example

112

 

17"”“1

 

DDE, can mimic the effects of sex hormones, are known as endocrine disruptors
and have been associated with asthma. (20) One of the findings in the first
research paper of this dissertation was the augmenting effect of female sex on
the interaction between maternal smoking and IL13 gene on CS-lgE. This
observation highlights the importance of hormones and the environmental
pollutants that mimic their physiological effects (endocrine disruptors) in asthma.
Therefore it is suggested that future studies investigate the gene-environment
interaction for pollutants other than tobacco smoke.

In summary, it is of public health importance to consider the prenatal period
as a critical time in the pathogenesis of diseases and to begin education
regarding exposure prevention before a pregnancy occurs. Future studies may
explore the molecular basis for the interaction between prenatal exposure to
tobacco smoke (and other pollutants) and genes. If epigenetic changes play a
role in the pathogenesis of asthma, it may even be possible to reverse the effect

of pollutants through intervention therapies.

113

APPENDIX

114

The programs and procedures used to perform the statistical analysis are
presented here. This appendix has three sections, HAPLOVIEW 3.2, PHASE
2.0.2 and SAS 9.0.1 programs used to perform statistical analysis.
HAPLOVIEW 3.2

To run HAPLOVIEW two datasets are needed:

First: A flat file data set that provides the program with the name of SNPs

and their corresponding position, saved as *.info.

r31800925 132069025
rs2066960 132070651
rsl295686 132072059
r820541 132072180
r81295685 132072661

Second: A flat file data set that provides the program with the identification
number (ID) of each individual and genotypes for each SNP. The first two
columns are IDs. Columns 3-6 have are constantly ‘0 0 1 1’. Each pair of
subsequent columns (5 pairs here as there are 5 SNPs) are genotype
information for the corresponding SNP. This file should be saved as *.ped and

this is how it looks like:

1100 1100 0000 0000
2200 1122 2222 0033
3300 1122 2222 0033

1534153400 1100 0000 0000
1535153500 1124 2222 0033
1536153600 1124 0000 0033

For genotype information (columns 7-16) ‘0’ stand for missing, ‘1’ stands
for ‘A’, ‘2’ stands for ‘C’, ‘3’ stands for ‘G’ and ‘4’ stands for ‘T’. To run the
program the option ‘Load genotype” should be chosen. The software browses the

two data sets and runs.

115

PHASE 2.0.2

Step 1: A SAS program to get Phase input.

libname cohort89 '1:\IOW\data';
data IL13_ALLELEI; set cohort89.i113_iow_v15;
keep id C15862743_1 C8932056_1 C8932052_1;

if substr(C15862743_10,1,1) = 'A' then C15862743_1 = 'A';

if substr(C15862743_10,1,1) 'C' then C15862743_1 = 'C';

if substr(C15862743_10,1,1) IN (", '-') then C15862743_1 = 'X';
if substr(C8932056_10,1,1) = 'T' then C8932056_1 = 'T';

if substr(C8932056_10,1,1) = 'C' then C8932056_1 = 'C';

if substr(C8932056_10,1,1) IN (", '-'1 then C8932056_1 = 'X';
if substr(C8932052_10,1,1) = 'A‘ then C8932052_1 = 'A';

if substr(C8932052_10,1,1) = 'G' then C8932052_1 = 'G';

if substr(C8932052_10,1,1) IN (", '-') then C8932052_1 = 'X';
where C15862743_10 not in (", '-') or

C8932052_10 not in (", '-') or C8932056_10 not in (", '-'),
run;

data IL13_ALLELEZ; set cohort89.i113_iow;y15;

keep id C15862743_2 C8932056_2 C8932052_2;

if substr(C15862743_10,2,1) = 'A' then C15862743_2 = 'A';

if substr(C15862743_10,2,1) = 'C' then C15862743_2 = 'C';

if substr(C15862743_10,2,1) IN (", '-') then C15862743_2 = 'X';
if substr(C8932056_10,2,1) = 'T' then C8932056_2 = 'T';

if substr(C8932056_10,2,1) = 'C' then C8932056_2 = 'C';

if substr(C8932056_10,2,1) IN (", '-') then C8932056_2 = 'X';
if substr(C8932052_10,2,1) = 'A' then C8932052_2 = 'A';

if substr(C8932052_10,2,1) = 'G' then C8932052_2 = 'G';

if substr(C8932052_10,2,1) IN (", '-') then C8932052_2 = 'X';
where C15862743_10 not in (", '—') or

C8932056_10 not in (", '-') or C8932052_10 not in (", '-');
RUN;

DATA IL13_ALLELE;MERGE IL13_ALLELE1 IL13_ALLELE2;BY ID;RUN;

DATA A;SET IL13_ALLELE;

FIL
OPT
PUT
@1

/
TIT

E PRINT NOTITLES;
IONS NODATE;

ID @6 C15862743_1 @8 C8932056_1 @10 ca932052_1
@6 C15862743_2 @8 C8932056_2 @10 C8932052_2;

LE;

RUN;

116

Step 2: input file for PHASE 2.0.2

In order to have the input file for PHASE 2.0.2 the output of the program
from step 1 was saved as a word file. Then missing values (X) were replaced by
question marks (?). Next step was to type the number of observations (907), loci

(3) and their type (SSS). The file should be saved as *.inp. This is the look for the

PHASEinput:

907

3

sss

2 CCG
CCG

3 CCG
CCG

4 ACA

1535 c c G
CTG

1536 ? c G
?TG

Step 3: Running PHASE 2.0.2. in DOS environment

<5 Command Prompt

Microsoft windows XP [Uersion 5.1.2600]
C) Copyright 1985-2001 Microsoft Corp.

C:\)1:

L \>cd iou\ill3\phase

L:\IOU\ilI3\Phase>phase phase_ill3_435256.inp ill3_435258.out_

 

This run produced several output files. | used ‘il13_435256.out_freqs’ to

obtain the results for estimated haplotype frequencies.

117

Step 4:
The file ‘il13_435256.out_pairs’ from previous step is used to extract
haplotype pairs with the highest probability for each individual and then merge

this information into the original data.

data in3;
infile 'L:\Iow\i113\phase\i113_435256.out_pairs' dsd truncover;
retain last ;
/* check to see what kind of line we're reading in
and act accordingly */
input dummy $1-3 @;
if dummy='IND' then do;
input @1 IND :$10. @;

last=IND;
end;
else do;
input @1 (varsl—vars3) (:$20.) ;
var=trim(varsl)||"_"||varsz;
end;
/* pull just the numeric off the IND value */
IND=scan(last, 2, ':');
if vars3=' ' then delete;
drop last dummy varsl varsz;

run;
proc sort; by ind;

/* collapse our by groups so we have 1 line per
IND value */
data cohort89.i113_435256;
set in3;
by ind;
retain a1-a5 b1-b5;
array a[5] ;
array b[S] $9. ;
if first.ind then do;

i=1;
do j=1 to dim(a);
alj]=' ';
b[j]=' 'i
end;
end;

a[i]=vars3;
b[i]=var;
if last.ind then output;
if i<5 then i+1;
drop varsB var i j;run;

proc print ; where b5 ne ";run;

data max;set cohort89.i113_435256;prob_best=max(a1,a2,a3,a4,a5);length
id 5.;run;

data cohort89.i113_435256snps(keep: id i113_diplo435256);

118

set max;

id=ind;

if a1=prob_best then ill3_diplo435256=b1;
if a2=prob_best then ill3_diplo435256=b2;
if a3=prob_best then i113_diplo435256=b3;
if a4=prob_best then i113_diplo435256=b4;
if a5=prob_best then i113_diplo435256=b5;
run;

**Making version 16;

proc sort data=cohort89.i113_435256snps;by id;run;
proc sort data=cohort89.i113_iow_v15;by id;run;

data cohort89.i113_iow_v16;merge cohort89.ill3_iow_v15
cohort89.i113_4352563nps;by id;

run;

119

SAS programs used to perform statistical analysis

libname cohort89 '1:\IOW\data';

data templ; set cohort89.i113_iow_v16;
IF C8932056_10 IN ('TT','CT') THEN C8932056='TX';
IF C8932056_10 IN ('CC') THEN C8932056='CC';

IF C8932052_10 IN ('AA','AG') THEN C8932052='XA';
IF C8932052_10 IN ('GG') THEN C8932052='GG';

IF C8932053_10 IN ('TT','CT') THEN C8932053='TX';
IF C8932053_10 IN ('CC') THEN C8932053='CC';

IF C15862743_10 IN ('AA','AC') THEN C15862743='AX';
IF C15862743_10 IN ('CC') THEN C15862743='CC';

if ASCON_1 =1 and ASCON_2 = 1 then asthma12= 1;
else if ASCON_1 ne . or ASCON_2 ne . then asthma12= 0;

age_mobi:(childdob-madob03)/365.25;
if 1< age_mobi < 24 then gr_age_mobi = 1;

if 24 <= age_mobi < 34 then gr_age_mobi 2;
if 34 <= age_mobi then gr_age_mobi = 3;
pcf2 = pcf3;

if pcf2 > 2 then pcf2 = 2;

if cb_ige=. then cige=.;
if 0<cb_ige<.5 then cige=0;
if .5=<cb_ige then cige=1;

if Delivery=1 then noninst_del=1;
else if Delivery ne . then noninst_del=0;

if neoncomp in (1,2,3,4,5) then ncomp=1;
if neoncomp in(9,99,100,.) then ncomp=0;

if ANTECOMP ne . then acomp=1;
if ANTECOMP=. then acomp=0;

if cat_0=1 or dog_0=1 then catdogo=1;
if cat_0=0 and dog_0=0 then catdogo=0;

if spt_48=1 or spt_10=1 then spt=1;
if spt_48=0 and spt_10=0 then spt=0;

if mb in(3,4,5) then season = 'spring';
if mb in(6,7,8) then season = 'summer';
if mb in(9,10,11) then season = 'fall';
if mb in(12,1,2) then season = 'winter';

if msmk_03=0 and (etsl=1 or etsz=1 or ets4=1 or anysmk=1)
then aft_p_effect='not in p but after';

120

if msmk_03=1 then aft_p_effect='in p, after y/n';

if msmk_03=0 and (etsl ne 1 and etsz ne 1 and ets4 ne 1 and anysmk ne
1) and

(etsl=0 or etsZ=0 or ets4=0 or anysmk=0) then aft_p_effect='xneither';

if aft_p_effect='not in p but after' then smk_after=1;
else if aft_p_effect ne " then smk_after=0;

if aft_p_effect='in p, after y/n' then smk_during=1;
else if aft_p_effect ne " then smk_during=0;

if (C8932056_10 NOT IN (",'-')or C15862743_10 NOT IN (",'-')Or
C8932053_10 NOT IN (",'-')or C8932052_10 NOT IN (",'-')) and CdaIO ne
. then miss=0;

if i113_diplo435256 in (

'ACA;ATA',

'ACG_ACA',

'ACG_ACG',

'ACG_ATA',

'ATG_ATA',

'CCA;ACA',

'CCA;CCA',

'CCA;CTA',

'CTA;ACA',

'CTA‘ATA',

'CTA_CTA',

'CTG_ATA',

'CTG_CTA',

'CTG_CTG') then minor_minor=1; else if i113_diplo435256 ne " then
minor_minor=0;

if ill3_diplo435256 in (

'CCG_ACA',

'CCG_ACG',

'CCG_ATA',

'CCG_ATG',

'CCG_CCA',

'CCG_CTA',

'CCG_CTG')then CCG_minor=1; else if i113_diplo435256 ne " then
CCG_minor=0;

if i113_diplo435256 in ('CCG_CCG')then homo_z=1;
else if i113_diplo435256 ne " then homo_z=0;
else if HETE_Z=. OR HOMO_Z=. THEN homo_m=.;

if (C15862743_10 NOT IN (",'-') and C8932056_10 NOT IN (",'-') and
C8932052_10 NOT IN (",'~')) and aft_p_effect ne " then miss=0;
run;

*****Table3.1.1;

proc freq data=temp1;tables
gr_age_mobi

mum_03

dad_03

msmk_03

121

dog_O

cat_o

season

boy

pcf2

acomp
noninst_del
ncomp

cige

/1ist missing;
where cige ne .;
run;

proc freq data=temp1;tables
gr_age_mobi
mum_03

dad_03

msmk_03

dog_o

cat_o

season

boy

pcf2

acomp
noninst_del
ncomp

cige

/list missing;
where miss=0 and cige ne .;
run;

*****Table 3.1.2;

proc traq data=temp1;

tables cige*

(gr_age_mobi

mum_03

dad_03

msmk_03

dog_O

cat_O

season

boy

pcf2

acomp

noninst_del

ncomp);

run;

proc logistic data=temp1 descending;
class gr_age_mobi ;

model cige=gr_age_mobi /r1 ;

run;

proc logistic data=temp1 descending;
model cige=mum_03 /rl ;

run;

proc logistic data=temp1 descending;
model cige=dad_03 /rl ;

run;

proc logistic data=temp1 descending;

122

model cige=msmk_03/rl ;

run;

proc logistic data=temp1 descending;
model cige=dog_0 /rl ;

run;

proc logistic data=temp1 descending;
model cige=cat_0/rl ;

run;

proc logistic data=temp1 descending;
class season ;

model cige=season /rl ;

run;

proc logistic data=temp1 descending;
model cige=boy /rl ;

run;

proc logistic data=temp1 descending;
model cige=pcf2 /rl ;

run;

proc logistic data=temp1 descending;
model cige=acomp /rl ;

run;

proc logistic data=temp1 descending;
model cige=noninst_del /rl ;

run;

proc logistic data=temp1 descending;
model cige=ncomp /rl ;

run;

**tab1e 3.1.3*;

proc freq data=temp1;

tables C8932056_10 /list;

where C8932056_10 not in (",'-');run;
proc freq data=temp1;

tables C15862743_10 /list;

where C15862743_10 not in (",'-');run;
proc freq data=temp1;

tables C8932053_10 /1ist;

where C8932053_10 not in (",'-');run;
proc freq data=temp1;

tables r130q /list;

where r130q not in (",'-');run;
proc freq data=temp1;

tables C8932052_10/list;

where C8932052_10 not in (",'-');run;

**tab1e 3.1.4*;

proc freq data=temp1;

tables ccg_minor*minor_minor*homo_z/list ;
run;

**Figure 3.1.1*;

proc freq data=temp1;

tables cige*homo_z/chisq;

tables msmk_03*cige*homo_z/chisq;
tables boy*msmk_03*cige*homo_z/chisq;
run;

**Figure 3.1.2*; .

proc logistic data=temp1 descending;

123

model cige=homo_z /rl ;

run;

proc logistic data=temp1 descending;
model cige=msmk_03 /rl ;

run;

proc logistic data=temp1 descending;
model cige=homo_z /rl ;

where msmk_03=0;

run;

proc logistic data=temp1 descending;
model cige=homo_z /rl ;

where msmk_03=1;

run;

*****Table3.2.1;

proc freq data=temp1;tables
(boy

bwt_03

ncomp

catdogo

season

gr_age_mobi

msmk_03

mum_03

dad_03)*cige;

where cige ne .;

run;

*****Table3.2.2;

proc freq data=temp1;tables

cige;

where cige ne .;

run;

proc freq data=temp1;tables
cige;

where cige ne . and asthmalz ne .;
run;

proc freq data=temp1;tables
cige;

where cige ne . and ascon_4 ne .;
run;

proc freq data=temp1;tables
cige;

where cige ne . and spt_48 ne .;
run;

proc freq data=temp1;tables
cige;

where cige ne . and cda10 ne .;
run;

proc freq data=temp1;tables
cige;

where cige ne . and spt_10 ne .;
run;

**Table 3.2.4*;

proc freq data=temp1;tables

(spt_48 spthdm48 sptgrs48 sptcat48 sptdog48 sptalt48 sptc1a48
spt_10 spthdmlo sptgralo sptcatlo sptdoglo sptaltlo sptcla10)*cige;
run;

124

proc logistic data=temp1 descending;

model cige=spt_48 /rl ;

title 'unadjusted LR';

run;

proc logistic data=temp1 descending;

model cige=spt_10 /rl ;

title 'unadjusted LR';run;

proc logistic data=temp1 descending;

class season gr_age_mobi;

model cige=boy msmk_03 gr_age_mobi mum_03 dad_03 season bwt_03 ncomp
catdogo spt_48 /rl ;

title 'Adjusted LR';run;

proc logistic data=temp1 descending;

class season gr_age_mobi;

model cige=boy msmk_03 gr_age_mobi mum_03 dad_03 season bwt_03 ncomp
catdogo spt_10 /r1 ;

title 'Adjusted LR';run;

**Table 3.2.5 and Figure 3.2.1*;
proc freq data=temp1;tables
asthma12*cige;

run;

proc logistic data=temp1 descending;
model asthma12=cige /rl ;

title 'Undjusted LR';run;

proc logistic data=temp1 descending;
model asthma12=boy msmk_03 gr_age_mobi mum_03 dad_03 season bwt_03
ncomp catdogo cige /r1 ;

title 'Adjusted LR';run;

run;

proc freq data=temp1;tables

ascon_4*cige;

run;

proc logistic data=temp1 descending;

model ascon_4=cige /r1 ;

title 'Undjusted LR';run;

proc logistic data=temp1 descending;

model ascon_4=boy msmk_03 gr_age_mobi mum_03 dad_03 season bwt_03 ncomp
catdogo cige /r1 ;

title 'Adjusted LR';run;

run;

proc sort data temp=1;by spt;run;

proc logistic data=temp1 descending;

model ascon_4=cige /r1 ;

title 'Undjusted LR';by spt;run;

proc logistic data=temp1 descending;

model ascon_4=boy msmk_03 gr_age_mobi mum_03 dad_03 season bwt_03 ncomp
catdogo cige /rl ;

title 'Adjusted LR';by spt;run;

run;

proc freq data=temp1;tables
cda10*cige;

run;

proc logistic data=temp1 descending;
model cda10=cige /rl ;

125

title 'Undjusted LR';run;

proc logistic data=temp1 descending;

model cda10=boy msmk_03 gr_age_mobi mum_03 dad_03 season bwt_03 ncomp
catdogo cige /rl ;

title 'Adjusted LR';run;

run;

proc sort data temp=1;by spt;run;

proc logistic data=temp1 descending;

model cda10=cige /rl ;

title 'Undjusted LR';by spt;run;

proc logistic data=temp1 descending;

model cda10=boy msmk_03 gr_age_mobi mum_03 dad_03 season bwt_03 ncomp
catdogo cige /rl ;

title 'Adjusted LR';by spt;run;

run;

**Figure 3.2.2*;

proc freq data=temp1;

table cige ;

title 'this calculation is presented as an example';
where asthma12 ne . ascon_4 ne . and cda10 ne .;run;

**Table 3.3.1***;

proc freq data=temp1;tables
phenoty

bwt_03

aft_p_effect

boy

cat_O

dog_O

/1ist missing;;

*where whezwhiz ne .;

where whezwhi2 ne . and phenoty ne . and aft_p_effect ne " and
miss=0;run;

data bias;

input level $ miss 5 n;

datalines;

a no 547

a yes 423

b no 239

b yes 156

run;

proc freq;

tables level*miss/chisq;

weight n;

run;

**Table 3.3.2***;

proc freq data=temp1;

tables C8932056_10 /list;

where C8932056_10 not in (",'-');run;
proc freq data=temp1;

tables C15862743_10 /1ist;

where C15862743_10 not in (",'—');run;

126

proc freq data=temp1;

tables C8932053_10 /list;

where C8932053_10 not in (",'-');run;
proc freq data=temp1;

tables r130q /list;

where r130q not in (",'-');run;

proc freq data=temp1;

tables C8932052_10/1ist;

where C8932052_10 not in (",'-');run;
**Table 3.3.2***;

proc logistic data=temp1 descending;
model phenoty=boy homo_zlsmk_during homo_zlsmk_after brst3mth bwt_03
cat_O dog_o /rl ;

title 'Eop';where phenoty in (2,0);
run;

proc logistic data=temp1 descending;

model phenoty=boy homo_zlsmk_during homo_zlsmk_after brst3mth bwt_03
cat_o dog_O /rl ;

title 'Lop';where phenoty in (3,0);

run;

proc logistic data=temp1 descending;

model phenoty=boy homo_zlsmk_during homo_z|smk_after brst3mth bwt_03
cat_O dog_o /rl ;

title 'ET';where phenoty in (1,0);

run;

**Figure 3.3.2;

proc freq data=temp1;

tables (C8932056 C15862743 C8932052 homo_z)*aft_p-effect*phenoty;
title 'Eop';where phenoty in (2,0);

run;

proc freq data=temp1;

tables (C8932056 C15862743 C8932052 homo_z)*aft_p_effect*phenoty;
title 'Lop';where phenoty in (3,0);

run;

proc freq data=temp1;

tables (C8932056 C15862743 C8932052 homo_z)*aft_p_effect*phenoty;
title 'ET';where phenoty in (1,0);

run;

127

REFERENCES

128

10.

11.

12.

13.

MERCKMedicus. Updated 2006. Available at:
h_ttp://www.merckmedicus.com

National institutes of Health. National Heart LaBI. GLOBAL STRATEGY
FOR ASTHMA MANAGEMENT AND PREVENTION. Updated 2005.
Available at: h_ttp:/Iwww.ginasthma.com.

Johansson SG, Bieber T, Dahl R, et al. Revised nomenclature for allergy
for global use: Report of the Nomenclature Review Committee of the
World Allergy Organization, October 2003. J Allergy Clin Immunol
2004;113:832-6.

Humbert M, Menz G, Ying S, et al. The immunopathology of extrinsic
(atopic) and intrinsic (non-atopic) asthma: more similarities than
differences. Immunol Today 1999;20:528-33.

Bousquet J, Jacot W, Yssel H, Vignola AM, Humbert M. Epigenetic
inheritance of fetal genes in allergic asthma. Allergy 2004;59:138-47.

Vignola AM, La Grutta S, Chiappara G, Benkeder A, Bellia V, Bonsignore
G. Cellular network in airways inflammation and remodelling. Paediatr
Respir Rev 2002;3:41-6.

Baldwin L, Roche WR. Does remodelling of the ainivay wall precede
asthma? Paediatr Respir Rev 2002;3:315—20.

Jeffery P. Inflammation and remodeling in the adult and child with asthma.
Pediatr Pulmonol Suppl 2001;21:3-16.

Arshad SH, Kurukulaaratchy RJ, Fenn M, Waterhouse L, Matthews S.
Rhinitis in 10-year-old children and early life risk factors for its
development. Acta Paediatr 2002;91:1334-8.

Arshad SH, Tariq SM, Matthews S, Hakim E. Sensitization to common
allergens and its association with allergic disorders at age 4 years: a
whole population birth cohort study. Pediatrics 2001;108:E33.

Holloway JW, Beghe B, Holgate ST. The genetic basis of atopic asthma.
Clin Exp Allergy 1999;29:1023-32.

Wiesch DG, Meyers DA, Bleecker ER. Genetics of asthma. J Allergy Clin
Immunol 1999;104:895-901.

Wiesch DG, Meyers DA. Strategies for analyzing genotype-phenotype
relationships in asthma. J Allergy Clin Immunol 2000;105:8482-6.

129

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

Martinez FD. Development of wheezing disorders and asthma in
preschool children. Pediatrics 2002;109:362-7.

Kurukulaaratchy RJ, Fenn MH, Waterhouse LM, Matthews SM, Holgate
ST, Arshad SH. Characterization of wheezing phenotypes in the first 10
years of life. Clin Exp Allergy 2003;33:573-8.

Kurukulaaratchy RJ, Fenn M, Twiselton R, Matthews S, Arshad SH. The
prevalence of asthma and wheezing illnesses amongst 10-year-old
schoolchildren. Respir Med 2002;96:163-9.

Mannino DM, Homa DM, Pertowski CA, et al. Surveillance for asthma--
United States, 1960-1995. MMWR CDC Surveill Summ 1998;47:1-27.

National Heatt L, and Blood Institute. Asthma Statistics. Bethesda, MD:
National Institutes of Health, Public Health Service, 1999.

Shaheen SO. Changing patterns of childhood infection and the rise in
allergic disease. Clin Exp Allergy 1995;25:1034-7.

Sunyer J, Torrent M, Munoz-Ortiz L, et al. Prenatal
dichlorodiphenyldichloroethylene (DDE) and asthma in children. Environ
Health Perspect 2005;113:1787-90.

Richardson G, Eick S, Jones R. How is the indoor environment related to
asthma?: literature review. J Adv Nurs 2005;52:328—39.

Jaakkola JJ, Gissler M. Maternal smoking in pregnancy, fetal
development, and childhood asthma. Am J Public Health 2004;94:136-40.

Holt PG. Programming for responsiveness to environmental antigens that
trigger allergic respiratory disease in adulthood is initiated during the
perinatal pen'od. Environ Health Perspect 1998;106 Suppl 3:795-800.

Cookson W. The alliance of genes and environment in asthma and
allergy. Nature 1999;402:85—1 1.

Gilliland FD, Li YF, Peters JM. Effects of maternal smoking during
pregnancy and environmental tobacco smoke on asthma and wheezing in
children. Am J Respir Crit Care Med 2001;163:429-36.

London SJ, James Gauderrnan W, Avol E, Rappaport EB, Peters JM.

Family history and the risk of early-onset persistent, early-onset transient,
and late-onset asthma. Epidemiology 2001;12:577—83.

130

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

Jaakkola JJ, Jaakkola MS. Effects of environmental tobacco smoke on the
respiratory health of children. Scand J Work Environ Health 2002;28 Suppl
2:71-83.

Nafstad P, Kongerud J, Botten G, Hagen JA, Jaakkola JJ. The role of
passive smoking in the development of bronchial obstruction during the
first 2 years of life. Epidemiology 1997;82293-7.

Gold DR, Burge HA, Carey V, Milton DK, Platts-Mills T, Weiss ST.
Predictors of repeated wheeze in the first year of life: the relative roles of
cockroach, birth weight, acute lower respiratory illness, and maternal
smoking. Am J Respir Crit Care Med 1999;160:227-36.

Colilla S, Nicolae D, Pluzhnikov A, et al. Evidence for gene-environment
interactions in a linkage study of asthma and smoking exposure. J Allergy
Clin Immunol 2003;111:840-6.

Moshammer H, Hoek G, Luttmann-Gibson H, et al. Parental Smoking and
Lung Function in Children: an lntemational Study. Am J Respir Crit Care
Med 2006.

Holgate ST. Genetic and environmental interaction in allergy and asthma.
J Allergy Clin Immunol 1999;104:1139-46.

Cookson WO, Young RP, Sandford AJ, et al. Maternal inheritance of
atopic lgE responsiveness on chromosome 11q. Lancet 1992;340:381-4.

Cookson WO, Sharp PA, Faux JA, Hopkin JM. Linkage between
immunoglobulin E responses underlying asthma and rhinitis and
chromosome 1 1 q. Lancet 1989;1 :1292-5.

Wills-Karp M, Luyimbazi J, Xu X, et al. Interleukin-13: central mediator of
allergic asthma. Science 1998;282:2258-61.

Wills-Karp M, Chiaramonte M. Interleukin-13 in asthma. Curr Opin Pulm
Med 2003;9:21-7.

Laoukili J, Perret E, Willems T, et al. IL-13 alters mucociliary differentiation
and ciliary beating of human respiratory epithelial cells. J Clin Invest
2001;108:1817-24.

Zhu Z, Homer RJ, Wang 2, et al. Pulmonary expression of interleukin-13
causes inflammation, mucus hypersecretion, subepithelial fibrosis,
physiologic abnormalities, and eotaxin production. J Clin Invest
1999;103:779-88.

131

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

CDC. Genomics and disease prevention. Updated 2005. Availabe at:
http://www.cdc.gov/genomicL/defaulihtm.

Kabesch M, Hoefler C, Carr D, Leupold W, Weiland SK, von Mutius E.
Glutathione S transferase deficiency and passive smoking increase
childhood asthma. Thorax 2004;59:569-73.

Wang Z, Chen C, Niu T, et al. Association of asthma with beta(2)-
adrenergic receptor gene polymorphism and cigarette smoking. Am J
Respir Crit Care Med 2001;163:1404-9.

Vercelli D. Genetics, epigenetics. and the environment: switching,
buffering, releasing. J Allergy Clin Immunol 2004;113:381-6; quiz 387.

Jaenisch R, Bird A. Epigenetic regulation of gene expression: how the
genome integrates intrinsic and environmental signals. Nat Genet 2003;33
Suppl:245-54.

Skinner MK, Anway MD. Seminiferous cord formation and germ-cell
programming: epigenetic transgenerational actions of endocrine
disruptors. Ann N Y Acad Sci 2005;1061z18-32.

Drake AJ, Walker BR, Seckl JR. lntergenerational consequences of fetal
programming by in utero exposure to glucocorticoids in rats. Am J Physiol
Regul Integr Comp Physiol 2005;288:R34-8.

Strachan DP. Hay fever, hygiene, and household size. ij
1989;299:1259-60.

Strachan DP. Family size, infection and atopy: the first decade of the
“hygiene hypothesis". Thorax 2000;55 Suppl 1:82-10.

Ball TM, Castro-Rodriguez JA, Griffith KA, Holberg CJ, Martinez F D,
Wright AL. Siblings, day-care attendance, and the risk of asthma and
wheezing during childhood. N Engl J Med 2000;343:538-43.

von Mutius E, Martinez FD, Fritzsch C, Nicolai T, Reitmeir P, Thiemann
HH. Skin test reactivity and number of siblings. ij 1994;308:692-5.

Riedler J, Braun-Fahrlander C, Eder W, et al. Exposure to farming in early
life and development of asthma and allergy: a cross-sectional survey.
Lancet 2001;358:1 129-33.

Sheikh A, Strachan DP. The hygiene theory: fact or fiction? Curr Opin
Otolaryngol Head Neck Surg 2004;12:232-6.

132

52.

53.

55.

56.

57.

58.

59.

60.

61.

62.

63.

64.

Karrnaus W, Johnson CC. Invited commentary: Sibship effects and a call
for a comparative disease approach. Am J Epidemiol 2005;162:133-8;
discussion 139.

Holt PG, Macaubas C. Development of long-term tolerance versus
sensitisation to environmental allergens during the perinatal period.
Current Opinion In Immunology 1997;9:782-7.

Strachan DP. Is allergic disease programmed in early life? [editorial:
comment]. Clinical and Experimental Allergy 1994;24:603-5.

Bjorksten B, Kjellman NI. Perinatal environmental factors influencing the
development of allergy. Clinical and Experimental Allergy 1990;20 Suppl
3:3-8.

Barker DJ. In utero programming of chronic disease. Clin Sci (Lond)
1998:9521 15-28.

Ravelli AC, van der Meulen JH, Michels RP, et al. Glucose tolerance in
adults after prenatal exposure to famine. Lancet 1998;351:173-7.

Warner JA, Jones CA, Williams TJ, Warner JO. Maternal programming in
asthma and allergy. Clinical and Experimental Allergy 1998;5z35—8.

Piccinni MP, Maggi E, Romagnani S. Role of honnone-controlled T-cell
cytokines in the maintenance of pregnancy. Biochem Soc Trans
2000;28:212-5.

Jones AC, Miles EA, Warner JO, Colwell BM, Bryant TN, Warner JA. Fetal
peripheral blood mononuclear cell proliferative responses to mitogenic and
allergenic stimuli during gestation. Pediatr Allergy Immunol 1996;7:109-16.

Romagnani S, Parronchi P, D'Elios MM, et al. An update on human Th1
and Th2 cells. Int Arch Allergy Immunol 1997;113:153-6.

Gaunt G, Ramin K. Immunological tolerance of the human fetus. Am J
Perinatol 2001 ;18:299-312.

Margni RA, Zenclussen AC. During pregnancy, in the context of a Th2-
type cytokine profile, serum lL-6 levels might condition the quality of the
synthesized antibodies. Am J Reprod Immunol 2001;46:181-7.

Prescott SL, Macaubas C, Holt BJ, et al. Transplacental priming of the
human immune system to environmental allergens: universal skewing of
initial T cell responses toward the Th2 cytokine profile. Journal Of
Immunology 1 998:160:4730-7.

133

65.

66.

67.

68.

69.

70.

71.

72.

73.

74.

75.

Kondo N, Kobayashi Y, Shinoda S, et al. Reduced interferon gamma
production by antigen-stimulated cord blood mononuclear cells is a risk
factor of allergic disorders—6-year follow- up study. Clinical and
Experimental Allergy 1998;28:1340-4.

Kondo N, Kobayashi Y, Shinoda S, et al. Cord blood lymphocyte
responses to food antigens for the prediction of allergic disorders. Arch
Dis Child 1992;67:1003-7.

Van Duren-Schmidt K, Pichler J, Ebner C, et al. Prenatal contact with
inhalant allergens. Pediatr Res 1997;41:128-31.

Bergmann RL, Edenharter G, Bergmann KE, et al. Predictability of early
atopy by cord blood-lgE and parental history. Clinical and Experimental
Allergy 1997;27:752-60.

Furuhashi M, Sugiura K, Katsumata Y, Oda H, Murase Y, Imai N. Cord
blood lgE against milk and egg antigens. Biol Neonate 1997;72:210-5.

Burrows B, Martinez FD, Halonen M, Barbee RA, Cline MG. Association of
asthma with serum lgE levels and skin-test reactivity to allergens. N Engl J
Med 1989;320:271-7.

Sears MR, Burrows B, Flannery EM, Herbison GP, Hewitt CJ, Holdaway
MD. Relation between airway responsiveness and serum lgE in children
with asthma and in apparently normal children. N Engl J Med
1991;325:1067-71.

Graves PE, Kabesch M, Halonen M, et al. A cluster of seven tightly linked
polymorphisms in the lL-13 gene is associated with total serum IgE levels
in three populations of white children. J Allergy Clin Immunol
2000;105:506-13.

Howard TD, Whittaker PA, Zaiman AL, et al. Identification and association
of polymorphisms in the interleukin-13 gene with asthma and atopy in a
Dutch population. Am J Respir Cell Mol Biol 2001;25:377-84.

Chang JC, Liu CA, Chuang H, et al. Gender-limited association of
cytotoxic T-lymphocyte antigen-4 (CTLA-4) polymorphism with cord blood
IgE levels. Pediatr Allergy Immunol 2004;15:506-12.

Ownby DR, Johnson CC, Peterson EL. Maternal smoking does not

influence cord semm lgE or lgD concentrations. J Allergy Clin Immunol
1991;88:555-60.

134

76.

77.

78.

79.

80.

81.

82.

83.

84.

85.

86.

Oryszczyn MP, Godin J, Annesi I, Hellier G, Kauffmann F. In utero
exposure to parental smoking, cotinine measurements, and cord blood
lgE. J Allergy Clin Immunol 1991;87:1169-74.

Magnusson CG. Maternal smoking influences cord serum lgE and lgD
levels and increases the risk for subsequent infant allergy. J Allergy Clin
Immunol 1986;78:898-904.

Kaan A, Dimich-Ward H, Manfreda J, et al. Cord blood lgE: its
determinants and prediction of development of asthma and other allergic
disorders at 12 months. Ann Allergy Asthma Immunol 2000;84:37-42.

Gonzalez-Quintela A, Vidal C, Gude F. Alcohol-induced alterations in
serum immunoglobulin e (lgE) levels in human subjects. Front Biosci
2002;7ze234-44.

Bjerke T, Hedegaard M, Henriksen TB, Nielsen BW, Schiotz PO. Several
genetic and environmental factors influence cord blood lgE concentration.
Pediatr Allergy Immunol 1994:5z88-94.

Bergmann RL, Schulz J, Gunther S, et al. Determinants of cord-blood lgE
concentrations in 6401 German neonates. Allergy 1995;50:65-71.

Orgel HA, Hamburger RN, Bazaral M, et al. Development of IgE and
allergy in infancy. journal of allergy and clinical immunology 1975;56:296-
307.

Nambu M, Shintaku N, Ohta S. Relationship between cord blood level of
IgE specific for Dermatophagoides pteronyssinus and allergic
manifestations in infancy. Biol Neonate 2003;83:102-6.

Hattevig G, Kjellman B, Bjorksten B. Clinical symptoms and lgE responses
to common food proteins and inhalants in the first 7 years of life. clinical
allergy 1987;17:571—8.

Stazi MA, Sampogna F, Montagano G, Grandolfo ME, Couilliot MF,
Annesi-Maesano l. Early life factors related to clinical manifestations of
atopic disease but not to skin-prick test positivity in young children.
pediatric allergy and immunology 2002;13:105-12.

Kjellman NI, Croner S. Cord blood lgE determination for allergy prediction-

-a follow-up to seven years of age in 1,651 children. Annals of Allergy
1984;53:167-71.

135

87.

88.

89.

90.

91.

92.

93.

94.

95.

96.

97.

Karrnaus W, Arshad H, Mattes J. Does the sibling effect have its origin in
utero? Investigating birth order, cord blood immunoglobulin E
concentration, and allergic sensitization at age 4 years. Am J Epidemiol
2001;154:909-15.

Croner S, Kjellman NI, Eriksson B, Roth A. lgE screening in 1701 newborn
infants and the development of atopic disease during infancy. Archives Of
Disease In Childhood 1982;57:364-8.

Ruiz RG, Richards D, Kemeny DM, Price JF. Neonatal lgE: a poor screen
for atopic disease. Clinical and Experimental Allergy 1991;21:467-72.

Michel FB, Bousquet J, Greillier P, Robinet-Levy M, Coulomb Y.
Comparison of cord blood immunoglobulin E concentrations and maternal
allergy for the prediction of atopic diseases in infancy. journal of allergy
and clinical immunology 1980;65:422-30.

Hansen LG, Hsst A, Halken S, et al. Cord blood lgE. III. Prediction of IgE
high-response and allergy. A follow-up at the age of 18 months. Allergy
1992;47:404-10.

Eiriksson TH, Sigurgeirsson B, Ardal B, Sigfusson A, Valdimarsson H.
Cord blood IgE levels are influenced by gestational age but do not predict
allergic manifestations in infants. pediatric allergy and immunology
1994:5z5-10.

Edenharter G, Bergmann RL, Bergmann KE, et al. Cord blood-lgE as risk
factor and predictor for atopic diseases [see comments]. Clinical and
Experimental Allergy 1998;28:671—8.

Strimas JH, Chi DS. Significance of IgE level in amniotic fluid and cord
blood for the prediction of allergy. Annals of Allergy 1988;61:133-6.

Woolcock AJ, Peat JK. Evidence for the increase in asthma worldwide.
Ciba Found Symp 1997;206:122-34; discussion 134-9, 157-9.

Milner AD, Marsh MJ, Ingram DM, Fox GF, Susiva C. Effects of smoking
in pregnancy on neonatal lung function. Arch Dis Child Fetal Neonatal Ed
1999:802F8-14.

Lodrup Carlsen KC, Jaakkola JJ, Nafstad P, Carlsen KH. In utero

exposure to cigarette smoking influences lung function at birth. Eur Respir
J 1997;10:1774-9.

136

98.

99.

100.

101.

102.

103.

104.

105.

106.

107.

108.

Stick SM, Burton PR, Gurrin L, Sly PD, LeSouef PN. Effects of maternal
smoking during pregnancy and a family history of asthma on respiratory
function in newborn infants. Lancet 1996;348:1060-4.

Brown RW, Hanrahan JP, Castile RG, Tager IB. Effect of maternal
smoking during pregnancy on passive respiratory mechanics in early
infancy. Pediatr Pulmonol 1995;1 9123-8.

Hanrahan JP, Tager IB, Segal MR, et al. The effect of maternal smoking
during pregnancy on early infant lung function. Am Rev Respir Dis
1992;145:1129-35.

Young S, Le Souef PN, Geelhoed 60, Stick SM, Turner KJ, Landau LI.
The influence of a family history of asthma and parental smoking on
airway responsiveness in early infancy. N Engl J Med 1991;324:1168-73.

Hoo AF, Henschen M, Dezateux C, Costeloe K, Stocks J. Respiratory
function among preterm infants whose mothers smoked during pregnancy.
Am J Respir Crit Care Med 1998;158:700-5.

Weiss ST, Tager IB, Munoz A, Speizer FE. The relationship of respiratory
infections in early childhood to the occurrence of increased levels of
bronchial responsiveness and atopy. Am Rev Respir Dis 1985;131:573-8.

O'Connor GT, Weiss ST, Tager IB, Speizer FE. The effect of passive
smoking on pulmonary function and nonspecific bronchial responsiveness
in a population-based sample of children and young adults. Am Rev
Respir Dis 1987;135:800-4.

Noakes PS, Holt PG, Prescott SL. Maternal smoking in pregnancy alters
neonatal cytokine responses. Allergy 2003;58:1053-8.

van der Pouw Kraan TC, van Veen A, Boeije LC, et al. An lL-13 promoter
polymorphism associated with increased risk of allergic asthma. Genes
lmmun 1999;1:61-5.

Weiss ST, Tager IB, Speizer F E, Rosner B. Persistent wheeze. Its relation
to respiratory illness, cigarette smoking, and level of pulmonary function in
a population sample of children. Am Rev Respir Dis 1980;122:697-707.

Leung TF, Tang NL, Chan IH, Li AM, Ha G, Larn CW. A polymorphism in

the coding region of interleukin-13 gene is associated with atopy but not
asthma in Chinese children. Clin Exp Allergy 2001 :31 :1515-21.

137

109.

110.

111.

112.

113.

114.

115.

116.

117.

118.

119.

120.

Hakonarson H, Bjomsdottir US, Ostennann E, et al. Allelic frequencies
and patterns of single-nucleotide polymorphisms in candidate genes for
asthma and atopy in Iceland. Am J Respir Crit Care Med 2001;164:2036-
44.

Devereux G, Barker RN, Seaton A. Antenatal determinants of neonatal
immune responses to allergens. Clin Exp Allergy 2002;32:43-50.

Asher Ml, Keil U, Anderson HR, et al. lntemational Study of Asthma and
Allergies in Childhood (ISAAC): rationale and methods. european
respiratory journal 1995;8z483-91.

SAS Institute. SAS/STAT Software. In: Cary, NC: SAS Institute Inc. 2001.
Rosner B. Fundamental of Biostatistics: Duxbury/Thomson, 2006.

OakRidge National Laboratory. Human Genome Project Information.
Updated 2006. Available at:
http://wwwoml.gov/sci/techresources/Human Genome/fag/snpsshtml.

The SNP Consortium. Single Nucleotide Polymorphisms for Biomedical
Research. Updated 2006. Available at: mpzllsnpcshlorg.

Hill WG, Robertson A. The effect of linkage on limits to artificial selection.
Genet Res 1966;8z269-94.

Stephens M, Smith NJ, Donnelly P. A new statistical method for haplotype
reconstruction from population data. Am J Hum Genet 2001;68:978-89.

The lntemational Study of Asthma and Allergies in Childhood - Steering
Committee. Worldwide variation in prevalence of symptoms of asthma,
allergic rhinoconjunctivitis, and atopic eczema: ISAAC. The lntemational
Study of Asthma and Allergies in Childhood (ISAAC) Steering Committee.
Lancet 1998;351:1225-32.

Sadeghnejad A, Kannaus W, Davis 3, Kurukulaaratchy RJ, Matthews S,
Arshad SH. Raised cord serum immunoglobulin E increases the risk of
allergic sensitisation at ages 4 and 10 and asthma at age 10. Thorax
2004;59:936-42.

SNPper. The Children's Hospital Informatics Program (CHIP)

Bioinforrnatics Tools. Harvard school of medicine (Boston, MA),
2005.m://snpper.chip.org)

138

121.

122.

123.

124.

125.

126.

127.

128.

129.

130.

131.

132.

Damon A, Damon ST, Reed RB, Valadian I. Age at menarche of mothers
and daughters, with a note on accuracy of recall. Hum Biol 1969;41:160-
75.

Hide DW, Arshad SH, Twiselton R, Stevens M. Cord serum lgE: an
insensitive method for prediction of atopy. Clin Exp Allergy 1991 :21 :739-
43.

Arshad SH, Stevens M, Hide DW. The effect of genetic and environmental
factors on the prevalence of allergic disorders at the age of two years. Clin
Exp Allergy 1993;23:504-11.

Barrett JC, Fry B, Mailer J, Daly MJ. Haploview: analysis and visualization
of LD and haplotype maps. Bioinforrnatics 2005;21:263-5.

Stephens M, Donnelly P. A comparison of bayesian methods for haplotype
reconstruction from population genotype data. Am J Hum Genet
2003:73:1 162-9.

Pattaro C, Heinrich J, Werner M, Marco R, stt M. Association between
interleukin-1 receptor antagonist gene and asthma-related traits in a
German adult population. Allergy 2006;61:239-44.

Peters U, Chatterjee N, Yeager M, et al. Association of genetic variants in
the calcium-sensing receptor with risk of colorectal adenoma. Cancer
Epidemiol Biomarkers Prev 2004;13:2181-6.

Szklo M, Nieto FJ. Epidemiology: Beyound the Basics: Aspen Publishers,
1999.

Tariq SM, Arshad SH, Matthews SM, Hakim EA. Elevated cord serum lgE
increases the risk of aeroallergen sensitization without increasing
respiratory allergic symptoms in early childhood. Clin Exp Allergy
1999;29:1042-8.

Arshad H. Influence of genetic and environmental factors on the level of
lgE at birth. pediatric allergy and immunology 1992;3:79-83.

Halonen M, Stern D, Taussig LM, Wright A, Ray CG, Martinez FD. The
predictive relationship between serum IgE levels at birth and subsequent
incidences of lower respiratory illnesses and eczema in infants. American
Review of Respiratory Disease 1992;146:866—70.

Puthothu B, Krueger M, Forster J, Heinzmann A. Association between

Severe Respiratory Syncytial Virus Infection and IL13/IL4 Haplotypes. J
Infect Dis 2006;193:438-41.

139

 

133.

134.

135.

136.

137.

138.

139.

140.

141.

142.

143.

144.

Bottini N, Borgiani P, Otsu A, et al. lL-4 receptor alpha chain genetic
polymorphism and total lgE levels in the English population: two-locus
haplotypes are more informative than individual SNPS. Clin Genet
2002:61 2288-92.

Wilson JN, Rockett K, Jallow M, et al. Analysis of IL10 haplotypic
associations with severe malaria. Genes lmmun 2005;62462-6.

Heinzmann A, Mac XQ, Akaiwa M, et al. Genetic variants of lL-13
signalling and human asthma and atopy. Hum Mol Genet 2000;9:549-59.

Flores D'Arcais A, De Biase D, Riva P, Tito A, Crippa L, Mariani E.
[Influence of maternal smoking on cord serum lgE levels in newborns].
Minerva Pediatrica 1993;45:443-7.

Oryszczyn MP, Annesi-Maesano l, Campagna D, Sahuquillo J, Huel G,
Kauffmann F. Head circumference at birth and maternal factors related to
cord blood total lgE. Clinical and Experimental Allergy 1999;29:334-41.

Sears MR, Burrows B, Flannery EM, Herbison GP, Holdaway MD. Atopy
in childhood. I. Gender and allergen related risks for development of hay
fever and asthma [see comments]. Clinical and Experimental Allergy
1993;23:941-8.

Dezateux C, Stocks J. Lung development and early origins of childhood
respiratory illness. Br Med Bull 1997;53:40-57.

Li YF, Gilliland FD, Berhane K, et al. Effects of in utero and environmental
tobacco smoke exposure on lung function in boys and girls with and
without asthma. Am J Respir Crit Care Med 2000;162:2097-104.

Xu B, Jarvelin MR, Hartikainen AL, Pekkanen J. Maternal age at
menarche and atopy among offspring at the age of 31 years. Thorax
2000;55:691-3.

Forbes L. Asthma and atopy: endocrine or metabolic conditions? Thorax
2005;60:793-4.

Brooks K, Samms—Vaughan M, Karrnaus W. Are oral contraceptive use
and pregnancy complications risk factors for atopic disorders among
offspring? Pediatr Allergy Immunol 2004;15:487-96.

Boroditsky RS, Reyes Fl, Winter JS, Faiman C. Serum human chorionic

gonadotropin and progesterone patterns in the last trimester of pregnancy:
relationship to fetal sex. Am J Obstet Gynecol 1975;121:238-41.

140

145.

146.

147.

148.

149.

150.

151.

152.

153.

154.

155.

Penny R, Olambiwonnu O, Frasier SD. Measurement of human chorionic
gonadotropin (HCG) concentrations in paired maternal and cord sera
using an assay specific for the beta subunit of HCG. Pediatrics
1976;58:110-4.

Ellinwood WE, Resko JA. Sex differences in biologically active and
immunoreactive gonadotropins in the fetal circulation of rhesus monkeys.
Endocrinology 1 980:1072902-7.

Gol M, Altunyurt S, Cimrin D, Guclu S, Bagci M, Demir N. Different
maternal serum hCG levels in pregnant women with female and male
fetuses: does fetal hypophyseaInadrenaI—gonadal axis play a role? J
Perinat Med 2004;32:342-5.

Liu X, Beaty TH, Deindl P, et al. Associations between total serum IgE
levels and the 6 potentially functional variants within the genes IL4, IL13,
and IL4RA in German children: the German Multicenter Atopy Study. J
Allergy Clin Immunol 2003;1 12:382-8.

Avni O, Lee D, Macian F, Szabo SJ, Glimcher LH, Rao A. T(H) cell
differentiation is accompanied by dynamic changes in histone acetylation
of cytokine genes. Nat Immunol 2002;3:643-51.

Fletcher R. Clinical epidemiology, The essentials. Baltimore: William &
Wilkins, 1996.

Haus M, Heese HD, Weinberg EG, Potter PC, Malherbe D, Hall JM.
Genetic and environmental influences on cord blood serum lgE and on
atopic sensitisation in infancy. S Afr Med J 1990;77:7-13.

Magnusson CG. Cord serum lgE in relation to family history and as
predictor of atopic disease in early infancy. Allergy 1988;43:241-51.

Croner S, Kjellman NI. Development of atopic disease in relation to family
history and cord blood lgE levels. Eleven year follow up in 1654 children.
pediatric allergy and immunology 1990:1214-20.

Kuehr J, Frischer T, Kannaus W, et al. Early childhood risk factors for
sensitization at school age. journal of allergy and clinical immunology
1992;90:358-63.

Bodtger U, Jacobsen CR, Poulsen LK, Malling HJ. Long-term repeatability

of the skin prick test is high when supported by history or allergen-
sensitivity tests: a prospective clinical study. Allergy 2003;58:1180-6.

141

156.

157.

158.

159.

160.

161.

162.

163.

164.

165.

166.

167.

Szabo SJ, Jacobson NG, Dighe AS, Gubler U, Murphy KM.
Developmental commitment to the Th2 lineage by extinction of IL-12
signaling. Immunity 1995;22665-75.

Perez VL, Lederer JA, Lichtman AH, Abbas AK. Stability of Th1 and Th2
populations. Int Immunol 1995;7z869-75.

Murphy E, Shibuya K, Hosken N, et al. Reversibility of T helper 1 and 2
populations is lost after long-term stimulation. J Exp Med 1996;183:901-
13.

Kuehr J, Karrnaus W, Forster J, et al. Sensitization to four common
inhalant allergens within 302 nuclear families. Clinical and Experimental
Allergy 1993;23:600-5.

Johnson CC, Ownby DR, Peterson EL. Parental history of atopic disease
and concentration of cord blood lgE. Clinical and Experimental Allergy
1996;26:624-9.

Wright AL, Taussig LM. Lessons from long-term cohort studies. Childhood
asthma. Eur Respir J Suppl 1998;27:17s-223.

lIIi S, von Mutius E, Lau S, et al. The pattern of atopic sensitization is
associated with the development of asthma in childhood. journal of allergy
and clinical immunology 2001;108:709-14.

Kurukulaaratchy R, Fenn M, Matthews S, Hasan Arshad S. The
prevalence, characteristics of and early life risk factors for eczema in 10-
year-old children. Pediatr Allergy Immunol 2003;14:178-83.

Taussig LM, Wright AL, Holberg CJ, Halonen M, Morgan WJ, Martinez
FD. Tucson Children‘s Respiratory Study: 1980 to present. journal of
allergy and clinical immunology 2003;111:661-75; quiz 676.

Howard TD, Koppelman GH, Xu J, et al. Gene-gene interaction in asthma:
IL4RA and IL13 in a Dutch population with asthma. Am J Hum Genet
2002;70:230-6.

Kurukulaaratchy RJ, Waterhouse L, Matthews SM, Arshad SH. Are
influences during pregnancy associated with wheezing phenotypes during
the first decade of life? Acta Paediatr 2005;94:553-8.

Meyers DA. Genetics in asthma and allergy. Immunol Allergy Clin North
Am 2005;25:ix-x.

142

168.

169.

170.

171.

172.

173.

174.

175.

Hall IP, Blakey JD. Genetic association studies in Thorax. Thorax
2005;60:357-9.

Barker DJ, Record RG. The relationship of the presence of disease to
birth order and maternal age. Am J Hum Genet 1967;19:433-49.

Barker DJ, Gluckman PD, Robinson JS. Conference report: fetal origins of
adult disease—report of the First lntemational Study Group, Sydney, 29-30
October 1994. Placenta 1995;16:317-20.

Barker DJ. Fetal programming of coronary heart disease. Trends
Endocrinol Metab 2002;13:364-8.

Miller RL, Chew GL, Bell CA, et al. Prenatal exposure, maternal
sensitization, and sensitization in utero to indoor allergens in an inner-city
cohort. American Journal of Respiratory and Critical Care Medicine
2001;164:995-1001.

Sutherland JE, Costa M. Epigenetics and the environment. Ann N Y Acad
Sci 2003;983:151-60.

Gawronska-Szklarz B, Luszawska-Kutrzeba T, Czaja-Bulsa G, Kurzawski
G. Relationship between acetylation polymorphism and risk of atopic
diseases. clinical pharmacology and therapeutics 1999;65:562-9.

Moodie FM, Marwick JA, Anderson CS, et al. Oxidative stress and
cigarette smoke alter chromatin remodeling but differentially regulate NF-
kappaB activation and proinflammatory cytokine release in alveolar
epithelial cells. Faseb J 2004;18:1897-9.

143

MICH GAN

lIll]IIIIIIIIIIIIIIIIIIIIIIIIl

3 1293