CHILD Banner

PRIMARY HYPOTHESIS:

The development of childhood asthma is dependent on the interactions of genetic predisposition and critical prenatal and early childhood environments including physical and microbial environments, nutrition, infections and social environment (socioeconomic status and stress).

Work - in - Progress

WIPThe CHILD Study is an evolving project. When it was developed, a primary hypothesis was articulated and it remains a foundational constant for the research inquiry. However, the CHILD Study also developed a series of sub-hypotheses which will continue to evolve as new theories emerge or as more evidence emerges to support other ideas.

Listed below are the main sub-hypotheses that represent current interests and have shaped the methods and statistical approaches used in the project. Over the course of the study, the CHILD researchers will refine, add to, or remove some of the hypotheses. Therefore, the following offer a current state of thinking but are subject to ongoing review.

 

TABLE H_01: THE MAIN HYPOTHESIS OF THE CHILD STUDY AND 7 SUB-HYPOTHESES

 

The development of childhood asthma is dependent on the interactions of genetic predisposition and critical prenatal and early childhood environments including physical and microbial environments, nutrition, infections and social environment (socioeconomic status and stress).

 

1

Innate immunity

Early life innate immune responses determine the onset of adaptive immune driven atopy and asthma.
2

Nutrition and Intestinal Microbiome

Pre-natal and post-natal nutrition and intrapartum care affect the development of childhood asthma through changes in the developing immune system and intestinal tract microbiome.
3

Infant Lung Function and Infection

Development of asthma is directly related to lung function in early childhood and the impact of viral infections on the lung.
4

Psychosocial

Early childhood socioeconomic status, through life stressors, material deprivation, and inadequate living conditions, interacts with the physical environment and genotype to influence the development of asthma.
5

Exposures

The impact of the physical environment on the development of childhood allergy and asthma, through modulation of the developing immune system, can be explained by exposures that collectively contribute to chronic inflammation via oxidative stress pathways. The Physical Environment: Indoor Air: Multiple exposures have been associated with wheeze.
6

Genetics

Specific gene-environment interactions predict the development of allergic and asthmatic phenotypes.
7

Epigenetics

Epigenetic mechanisms mediate the effects of early life environments to persistently alter the expression of genes and functions of cells critical to the development of allergic diseases and asthma.

 

CHILD Logo and Tag

Early Life Innate Immune Responses Determine the Onset of Adaptive Immune Driven Atopy and Asthma.

 

Antigen-dependent adaptive reactions constitute the central core of the immune mediated pathogenesis underlying asthma. With an increase in mechanistic understanding also came the realization that innate responses directly precede and then direct adaptive responses in the pathogenesis of asthma. Innate immunity has thus become a highly attractive target not only for therapeutic but also prophylactic interventions and biomarker discovery in asthma and allergy.

 

The innate immune system constantly scans the environment employing a multitude of sensors, including pattern recognition receptors (PRR). The information gathered by these PRRs signals the adaptive immune system to remain either resting or become activated, and if activated to exert specific effector functions (e.g. Th1 vs. Th2 vs. Th17 vs. Treg etc.). There have already been surprising insights into innate immune sensing in relation to atopy and asthma. Several studies have identified that the risk of eczema appears associated with variation in cytokine production in cord blood in response to a variety of PRR stimuli. However, none of these studies applied the full force of a comprehensive innate immune analysis along with precise clinical assessments in the context of specific environmental and genetic variables that we propose here, and it is highly likely that important interactions were missed in past studies.

 

There are important qualitative and quantitative differences in age-related innate immune responses, which affect the response to allergic stimuli. These patterns of innate immune responses represent 'windows of vulnerability' to induction of allergic responses. We have developed a comprehensive and robust platform to capture these patterns in early life innate ontogeny using the panel of stimulations employed in CHILD.

The final correlation of the environmental variables measured in CHILD (on the background of particular host genetic make-up) with our innate immune developmental analysis and carefully diagnosed asthma and atopic disease represents an unprecedented opportunity for identifying not only underlying complex interactions and mechanisms, but also novel diagnostic and prognostic biomarkers, and even therapeutic interventions.

 

Specific testable hypotheses related to INNATE IMMUNITY:


  1. Pregnancy results in enhanced systemic innate (TLR, RLR, NLR) immune capacity Longitudinal study, n=400 women – already in progress using surplus CHILD blood in Winnipeg.
  2. Differences in innate immune reactivity a) precede and b) follow asthma and atopy i.e. are predictive and therapeutic biomarkers (also connected with microbiome: mechanistic). Case control study using innate plates from cord blood, 1 year blood and 5 year blood.
  3. Innate immune readouts correlate with outcome measures of atopic/asthmatic state. This reflects the usefulness of collecting longitudinal data to look at not only onset of atopy and asthma, but also potentially "outgrowing" the disorders (longitudinal biomarkers).

 

Specific testable hypotheses related to ADAPTIVE IMMUNITY:


A number of hypotheses will be tested by stimulating cord blood mononuclear cells (CBMCs) and/or peripheral blood mononuclear cells (PBMCs) and measuring the cytokine responses.

  1. Sensitization to egg or peanut at 1 year is a major risk factor for wheezing episodes in early infancy. Timeline 2013-2014. A possible synergy between CHILD and the CALIPER project is being explored which may provide these data.
  2. A Th2 weighted immune response (vs. Th1 or Th17) is associated with the development of asthma and atopy. To address this, CBMCs will be stimulated with ionomycin/IL2 and cytokine responses documented by flow cytometry. Cytokine responses will be correlated with subsequent wheezing and asthma diagnosis. For economy of samples and costs, these analyses are best deferred until the 5 year outcome is known, when a case-control approach will be possible.
  3. The adaptive immune response to specific allergens is associated with the development of asthma and atopy. CBMCs will be stimulated with allergen and cytokine responses documented by FACS.
  4. The adaptive immune response to specific viruses is associated with the development of asthma and atopy. CBMCs will be stimulated with viral proteins cytokine responses documented by FACS.
  5. Sensitization to egg or peanut at 1 year is a major risk factor for persistent asthma at age 5 years.

 

NOVELTY of the CHILD study:

Given the age group most affected by atopic disease (infants and young children), blood sample volumes obtained from subjects in studies of asthma will be small. Analysis of the immune development in relation to atopy and asthma will therefore require a multiplexed platform to extract the most information from the smallest samples. Analysis of the innate immune response has been complicated by the fact that the innate immune system has a strong sentinel (environment sensing) function, i.e. is highly susceptible to even the smallest variation in procedure. Furthermore, if analysis of the innate immune responses is to be meaningful, it requires capturing their complex, multifaceted interactions (the "interactome"). With a focus on early life immune responses, we have developed a robust platform that allows high-throughput, multiplexed innate immune analysis of the smallest of human samples (Kollmann 2009; Corbett 2010; Lavoie 2010). This platform includes:


  1. The necessary infrastructure (we have use of one of only 2 LSRII special order flow cytometers currently in Canada, allowing assessment of up to 21 parameters per cell simultaneously at up to 3000 cells/sec; we also have state-of-the art Luminex based multiplexing capacity; and we have an advanced Beckman robotic system promoting high-throughput, standardized processing of samples);
  2. Stringent standard operating procedures (SOPs) and QC/QA for processing (Jansen 2008; Fortuno 2011) as well as analysis (Lee 2008, Blimkie 2010; Fortuno 2010; Spidlen 2011);
  3. The necessary bioinformatic tools to capitalize on the rich, complex data generated (Mookherjee 2009; Gardy 2009; Ho 2010; Lynn 2010).

INNATE IMMUNE FUNCTION Power and Sample Size

Because of fiscal and practical constraints, innate immune function tests will be performed in the 500 children in Toronto undertaking infant pulmonary function testing, and in 500 of the 1,500 Vancouver children in the Stress cohort. All 1000 will have all other data points (genetics, environmental exposure etc.) collected as outlined in CHILD.

 

We expect that ~10% of the enrolled subjects will have physician diagnosed asthma by age 5, i.e. ~100 of the entire 1000 strong cohort with innate data, of which half will have early life PFT data as well. We also expect an additional ~20% of these children to suffer from allergy (diagnosed via skin prick and/or allergen specific IgE) or eczema (diagnosed clinically), i.e. an additional 200 atopic subjects for whom we have complete innate immune data will be classified as atopic.

 

We propose to match this group of 100 and 200 (i.e. a total of ~300) cases 1:2 with a non-asthmatic/non-atopic control group. Thus, the basic statistical study design is that of a stratified prospective cohort study with a binary outcome (yes or no asthma or atopy at age 5). Standard statistical models (e.g. logistic regression) will be used to relate innate immune data obtained at birth, 1 and 5 years of life to the odds of developing asthma/atopy at age 5. Statistical tests of primary hypotheses (e.g. dependence of risk to develop asthma/atopy on innate immune responses) will be performed at the standard 0.05 level of statistical significance. Because the innate immune baseline covariates can all be considered continuous measurements, the power of these tests is estimated generically by the power of a test for a linear trend in risk across quintiles of a continuous covariate. With the expected ~ 100 asthmatic and ~200 atopic cases and the matched ~600 non-asthmatic and non-atopic controls, an odds ratio will be calculated (for the highest vs. lowest quintile of the covariate) to be detected with 90% power

 

Therefore, 100 asthmatic subjects from the sub-cohorts of Vancouver and Toronto will be matched with 200 non-asthmatic subjects to relate innate immune data obtained at birth, age 1, and age 5 to our primary outcome. Using conditional logistic regression1 with a Type I error rate of 5%, we will have 81% power to detect an odds ratio (OR) of 1.4 per standard deviation (SD) for differences in the innate immune measures, e.g. 2TNF-α, and over 90% power to detect an OR of 1.72 per SD.

INNATE IMMUNE RESPONSE Methods

 

The analyses will be largely exploratory and will proceed by testing individual innate immune response parameters for association with risk for asthma. Because of the multiplicity of these tests, an adjustment will be done to correct for multiple comparisons.

 

The result of this series of tests will be a list of immune parameters with reliable statistical evidence for association with risk for asthma. Multivariate logistic regression models will then be used to explore the joint effects and possible interactions amongst the various innate immune parameters on risk for asthma or atopy. Host genetics and environmental exposures will then be added to the best multivariate model to assess the extent to which the observed gradient between host vs. environment in risk for asthma and atopy can be explained by the measured innate immune response profile. Finally, the most promising (i.e. most significantly associated with asthma or atopy) innate parameters will be included in the stepwise analysis of the entire data set obtained in CHILD.

 

Specifically, we will assess innate immune sentinel function as response to Toll-like receptor (TLR) ligands, as production of signals aimed to regulate subsequent adaptive immune function (cytokines and costimulatory molecule expression) as described (Kollmann 2009; Corbett 2010; Lavoie 2010). To this end, heparinised blood exposed to TLR ligands will be assessed for intracellular cytokine production and surface expression of costimulatory molecules in dendritic cells and monocytes via single cell cytokine profiling using polychromatic flow cytometry (LSRII). We will also use global blood cytokine profiling (by Cytometric Bead Arrays) to capture cytokines and chemokines currently not assessable by flow cytometry (e.g. TGFb). This will be done on fresh blood at birth (cord blood), and at 1 and 5 years of age (peripheral blood). The stimulated samples are stored frozen, and analyzed in pools at a later time point in one centralized facility.

 

Innate Immune Effector Function:

In addition to a complete blood count for eosinophil numbers (done on fresh blood), we will measure eosinophil and mast cell granule proteins and cytokines (including LTC4 production), as well as functional IDO levels, on cryopreserved samples taken at birth, and at 1and 5 years of age. Besides B-cell serum immunoglobulin levels (including IgE, IgG4), several T-cell functional assays will be combined in one single Flow Cytometry assay with the following readouts:

  • Antigen/Allergen specific T cell proliferation;
  • Antigen/Allergen specific single cell cytokine (Th1/Th2) profiling;
  • Antigen/Allergen specific global cytokine profile; and Regulatory T cell and NK/T characterization. We expect to confirm and extend findings of previous birth cohorts that focused primarily on T cells (particularly low T cell IFN-gamma production).

This antigen/allergen specific assessment will allow us to determine when sensitization occurs, which has important public health implications in terms of advising mothers on, e.g., their diet during pregnancy. Furthermore, allergen specific immune function will be assessed in the context of the entire immune system and environmental milieu in each individual, determining whether children who develop allergy/asthma have a generally different immune response, or a selective allergic reaction to specific allergens. This will be accomplished by comparing allergen to vaccine antigens (Tetanus Toxoid, Pertussis Toxin; selected allergens will be identified through skin testing). Determination of allergen specific acquired immune function will be undertaken at birth, 1 and 5 years of age, using peripheral blood frozen leukocyte samples taken and stored frozen on all children. All samples from one individual will be run in the same experiment.

 

Pre-natal and post-natal nutrition and intrapartum care affect the development of childhood asthma through changes in the developing immune system and intestinal tract microbiome.

 

Research on diet and interventions to prevent atopic disease have focused on foods with antiinflammatory properties (e.g. n-3 fatty acids), antioxidants (vitamin E and zinc), and vitamin D. Recent meta-analyses suggest beneficial effects for pre-natal vitamins A, D, and E, zinc, fruits and vegetables, and the Mediterranean diet. Nutrients may impact development of asthma through immune modulation or the child's intestinal microbiome.

 

Microbiota composition during infancy, in particular, colonization with C. difficile, predicts recurrent wheeze or asthma later in childhood. Several interrelated exposures impact intestinal microbiota and are associated with allergic disease. Cesarean-section delivery interferes with the newborn's first microbial exposure, resulting in fewer intestinal Bifidobacteria and Bacteroides, and greater colonization with C. difficile.

 

As much as 20% higher rates of childhood asthma has been attributed to Cesarean-section compared with vaginal delivery. A new study has shown the association between Cesarean delivery and asthma development to be mediated by C. difficile. Antibiotics also affect intestinal colonization by suppressing commensal bacteria and causing the emergence of asthma-associated pathogens, such as C. difficile. antibiotic exposures whether in utero, during the neonatal period or infancy, or through breast-feeding have been associated with increased risk for wheeze or asthma later in childhood.

 

A statistically significant pooled estimate of this association has been found among studies adequately adjusting for respiratory infections. Two studies have found a stronger antibiotic effect in infants without prior genetic predisposition.

 

On the other hand, breast-feeding confers "beneficial" gut microbiota to infants. While new studies and meta-analyses continue to demonstrate that breastfeeding protects against recurrent wheeze and asthma in later childhood, these benefits may not apply to infants of atopic mothers. Breast milk of allergic mothers contains less beneficial bifidobacteria than non-allergic mothers, resulting in lower counts of fecal bifidobacteria in their infants.

 

Preliminary maternal data indicate substantial dietary deficiencies in iron and Vitamin D compared with that recommended during pregnancy (Table 2). Combining fetal and post-natal nutrition data will enable exploration of new hypotheses on vitamin D, intestinal microbiota and the developing immune system in infants.

 

Specific testable hypotheses related to NUTRITION and MICROBIOME:

  1. Child health outcomes are partly programmed in-utero, and influenced by maternal nutrition. The programmed fetus then transitions to, and interacts with, a new external nutritional environment which influences the microbiota. Collectively genetics, epigenetics, nutrition and microbiota impact childhood growth, adiposity related metabolic traits, allergic disorders and asthma. This is the primary hypothesis in the current application to CIHR Programmatic grants in Food and Health (LOI already approved). We will use DNA extracted from cord blood, and a new chip available by Illumina that is genome wide 500,000 SNPs -we propose doing this in as many CHILD cord blood and mothers as possible. Timeline 2-3 years.
  2. The intestinal microbiome is influenced by species present in the indoor environment where children spend the majority of their time. Analyses are in progress and will deliver results in the coming year. It may be a precursor to an "urban hygiene" hypothesis within CHILD, i.e. if microbes in the rural/farm environment are protective it is likely that there is a parallel for urban dwellers. Preliminary data have been presented, additional analyses supported through the GxE Platform project: BEAM.

SyMBIOTA Microbiome Project: Hypothesis: Antibiotic use, independent of and synergistically with formula-feeding and caesarean section delivery, is associated with altered fecal microbiota (diversity, prevalence) at age 3 months and 1 year. Changes in fecal microbiota are greater and more persistent following broad-spectrum antibiotic use. Work currently in progress, CIHR funded; ultimately dependent on access to the complete CHILD cohort to achieve required sample size. Secondary hypotheses include:


  • Compared to emergency caesarean section or vaginal delivery, planned caesarean delivery alters the colonization of the infant gut, leading to persistent changes
  • The composition and diversity of the infant gut is influenced by diet, including the type and relative amount of breastfeeding, formula feeding, and solid foods
  • Indirect antibiotic exposure, occurring via the mother during pregnancy or lactation, alters the infant gut microbiota.
  • The infant gut microbiota is influenced by environmental factors such as pets, siblings, daycare attendance, time spent outdoors and home cleaning practices
  • The infant gut microbiota is affected by parent occupation.
  • Infant use of gastroesophageal reflux disease drugs (GERDD) influences the composition and diversity of the gut microbiota.
  • Through their impact on the infant gut microbiota, antibiotics may increase the likelihood of childhood overweight.
  • Maternal overweight in pregnancy alters infant gut microbiome, increasing risk for child overweight
  • The composition and diversity of the gut microbiota is associated with changes to intestinal immunoglobulin A and circulatory markers of innate immune in 1 year old infants.

 

The NOVELTY of the CHILD study concerning the Infant Microbiome:

Microbial diversity plays a significant role in atopic diseases: children treated with antibiotics are at greater risk for asthma; probiotics have shown beneficial prophylactic effects for asthma; and gnotobiotic (microbiota-free) animals cannot generate tolerance to antigens. Another example is emerging evidence of breast-milk transmission of the maternal microbiome, potentially impacting the child immune system, and perhaps explaining some of the confusing and conflicting data surrounding the benefit, or lack of benefit, of breastfeeding on the risk of developing allergy and asthma.

 

New experimental approaches will be applied to examining meconium and stool, breast milk and nasal secretions in the CHILD Study. Maternal factors, including mode of delivery (vaginal versus caesarean section), antibiotic use, animal exposure, dietary factors (breast-feeding, introduction of table food) and other environmental influences, all impact the microbiome, both in the gut and airways. These microbiome studies will inform and guide the food and pharmaceutical industries with regard to healthy foods and novel, rational antimicrobial drugs and treatment regimens; public health professionals, regarding formulation of policies involving lifestyle choices for healthy maternal and infant diets and nutrition; and professional (pediatric and obstetric) management guidelines.

 

Preliminary Results - Microbiome

Our preliminary microbiome data indicate that breastfed infants are significantly less colonized with C difficile in their intestine, which will reduce the risk of atopic sensitization and wheeze at 1 year. Intestines of breastfed infants also had higher levels of Immunoglobulin A, which interacts with intestinal microbiota to train the infant's immune system. CHILD data on maternal diet during pregnancy, delivery method, infant feeding, antibiotic use, intestinal microbiota composition, immune function in infants and allergic disease outcomes will improve understanding of the role of breastfeeding and optimize potential interventions (e.g. probiotics) during pregnancy and infancy.

 

Preliminary Results - Nutrition

Data from the first 991 maternal Food Frequency Questionnaires have revealed a strikingly high prevalence (75-90%) of dietary deficiency for iron and Vitamin D, based on total dietary intakes, when compared with Estimated Average Requirements (EARs) currently recommended by Health Canada for Iron (http://www.hcsc. gc.ca/fn-an/nutrition/reference/ table/ref_elements_tbl-eng.php) and Vitamin D (http://www.hcsc. gc.ca/fn-an/nutrition/reference/table/ ref_vitam_tbl-eng.php). These findings have major public health implications now and in the future for Canadian mothers in pregnancy. CHILD will enable development of a quantitative risk assessment (analogous to dose-response) of these nutritional factors on health outcomes, a major goal of PHAC's Chronic Disease Management and Prevention Strategy.

 

Using a validated Food Frequency Questionnaire to examine maternal diet during pregnancy and regular questionnaire data about the infants' diets and their transition to table foods, analyses will focus on the role of prenatal food intake previously associated with the development of asthma and allergy. This includes Vitamin D, antioxidants (such as Vitamin E and zinc) and foods with purported anti-inflammatory properties such as omega-3 fatty acids. Novel analyses will examine the role of probiotics, nutraceuticals and alternative medicines (e.g., herbal teas) on atopy and on immune function development. In addition, the role of diet in altering the infant microbiome can be assessed.

 

The Food Frequency Questionnaire (FFQ) developed by the nutritional epidemiologists at the Fred Hutchinson Cancer Research Center in Seattle, WA and based on the questionnaires used in two large NIH funded studies, the Selenium and Vitamin E Cancer Prevention Trial (SELECT) and the VITamins and Lifestyle study (VITAL) was modified to reflect Canadian ethnic food choices. In addition, the database developed by the University of Minnesota Nutrition Data Systems for Research (NDSR) for data entry and nutrient analysis was "Canadianized" for use in CHILD. This questionnaire is a self-administered FFQ asking pregnant mothers to report the frequency of consumption and portion size of approximately 175 line items during their current pregnancy. These questionnaires are completed by the mother at the time of enrolment.

 

 

NUTRITION and MICROBIOME Power / Sample Size

A sample size of 3000 (including 300 asthmatics) will allow us to examine the relationship between maternal Vitamin D deficiency and offspring asthma3 with 80% power for an OR of 1.86, assuming 5% offspring asthma for mothers with adequate intake (or an OR of 1.60 assuming 10% offspring asthma).

MATERNAL NUTRITION Methods

Prenatal and postnatal nutrition are assessed in addition to specific measures related to breastfeeding and breast milk. A Food Frequency Questionnaire (FFQ) developed by the nutritional epidemiologists at the Fred Hutchinson Cancer Research Center in Seattle, WA and based on the questionnaires used in two large NIH funded studies, the Selenium and Vitamin E Cancer Prevention Trial (SELECT) and the VITamins and Lifestyle study (VITAL) was modified to reflect Canadian ethnic food choices. In addition, the database developed by the University of Minnesota Nutrition Data Systems for Research (NDSR) for data entry and nutrient analysis was "Canadianized" for use in CHILD. This questionnaire is a self-administered FFQ asking pregnant mothers to report the frequency of consumption and portion size of approximately 175 line items during their current pregnancy. These questionnaires are completed by the mother at the time of enrolment.

 

BREAST-FEEDING:

Breast-feeding status is assessed by questionnaire at 3, 6, 12 and 18 months of age. Items on the questionnaire include duration of exclusive breastfeeding, age of first supplementation, and the age of first introduction of solids.

 

BREAST MILK SAMPLE:

Breastfeeding mothers are asked to collect (by hand express or pump) and refrigerate 10 ml of breast milk in a collection tube within 24 hours of the 3-month home inspection visit, when it is collected and transported in a cooler with a cold pack to the central laboratory facility, centrifuged and the aqueous and lipid phases separated, aliquoted, and stored at -80oC for subsequent analyses. Hand expressed breast milk is collected from 1 day prior to the visit to include at least 2 separate feedings with milk collected before and after feeding. A total of 10 mls of breast is dispensed by study staff into 6 x 1.5 ml aliquots and frozen within 8 hours.

 

CHILDHOOD DIET:

Child diet will be assessed by a FFQ (currently being used by the FAMILY Study) at 3 and 5 years of age (Klohe et al, 2005; Iannotti et al, 1994). Anthropometric measures: Weight and length/height are recorded at birth (from the delivery chart), 3 months, 1 year, 3 years and 5 years of age.

 

MICROBIOME Methods

 

SAMPLE COLLECTION AND COMMUNITY DNA ISOLATION

Meconium and fecal samples will be collected from infants at 3 months of age and repeated at 1 yr of age. Samples consisting of 5–10 g stool (1–2 g for meconium) will be collected aseptically from a freshly soiled diaper using a sterile spatula, divided into aliquots and stored at –80oC in sterile containers. One frozen fecal sample will be thawed and be homogenized in phosphate buffered saline and the homogenate will be divided into two aliquots, one of which will be frozen for later analysis of fecal IgA. At regular intervals, an aliquot will be spiked with a known number of cells of a non-conflicting bacterium for use as a control in the evaluation of DNA isolation efficiency and the standardization of PCR. Total DNA will be isolated from this aliquot using a commercial kit (Qiagen QIAamp DNA Stool Mini kit) and an automated high throughput platform (QiaCube). DNA yield will be evaluated and standardized by agarose gel electrophoresis.

Community characterisation by qPCR: The abundance of selected bacterial taxa (e.g., Clostridium difficile, Bifidobacterium spp., Bacteroides fragilis) will be assessed by TaqMan PCR using described primer/probe combinations.

 

COMMUNITY CHARACTERIZATION BY HIGH-THROUGHPUT DNA SEQUENCING

A region spanning the V5, V6 & V7 subregions of the bacterial 16S rRNA gene will be amplified using universal primer sequences. Within this amplicon, 4 hypervariable 36 bp domains will be sequenced using the Metagenomic Analysis by Serial Illumina Sequencing (MASIS) method. This approach uses the Illumina Genome Analyzer (GA) II next-generation sequencing platform to generate multiple short, but phylogenetically informative reads from a single 16S rRNA template. Taxonomic classification of sequences will be generated by alignment against a bacterial reference database (e.g., Silva Non- Redundant 16S Ribosomal Database; Ribosomal Database Project, University of Michigan).

 

COMMUNITY COMPOSITION Analysis

Diversity indices and species richness estimators will be calculated using the R and QIIME software packages. Since the appropriateness of the particular measure of species richness is dependent on the dataset, a variety of methods will be used to estimate species richness including tests from the Chao and jackknife families of nonparametric methods as well as well as species-accumulation curve estimators. Diversity will be estimated using both the Shannon and Simpson indices. In addition to describing the community complexity, these statistics will also allow us to use a pilot set of samples to assess the completeness of sampling, and to determine the degree of sequence multiplexing that is needed for the sample set. Multivariate analyses will be performed to test for differences in community composition. Here we will use nonparametric methods such as analysis of similarity (ANOSIM) and permutational multivariate analysis of variance using distance matrices (ADONIS) with distance matrices calculated both from frequency and presence-absence datasets. Principal components analyses (PCA), incorporating exposure data, will also be performed based on both presence-absence and frequency of Operational Taxonomic Units (OTUs) .

 

ANTIBIOTIC EXPOSURES Measurement

Antibiotic and other prescription medication records for infants from provincial databases and data collection instruments used in the cohort. To protect patient confidentiality, all prescription records in provincial databases are anonymized and contain encrypted versions of the health identification number. Linkage between the prescription data, microbiota profiles and survey data from the CHILD study will be achieved through a CHILD identification number, which has been attached to the encrypted health identification number. The child's health identification number will be encrypted by the respective provincial health departments for data linkage purposes. Parent permission for this data linkage and the child's health identification number has been obtained on the consent form. The wording on the consent from to request this access was approved by the Human Research Ethics boards at the site-specific universities. Province-specific processes will be followed to receive approval for database access, to achieve the encryption of the health identification number and to obtain the prescription datasets. Other environmental exposure measures: Variables will be created to measure the events around birth including mode of delivery (vaginal unassisted, forceps and extraction; elective or emergency elective Caesarean section), premature rupture of membranes, gestational age, birth weight and type of feeding (breast, formula or both) in hospital are extracted in the CHILD study from the birth record. Maternal use of antibiotics and other medications (e.g. probiotic supplements) and diet during pregnancy are determined from questionnaires during pregnancy at 18 and 36 weeks. During the home visit at 3 months of age and clinic visit at one year, a child health questionnaire is administered to inquire about type and duration of feeding (breast, formula, both), and if applicable, the name of the formula and amount used to supplement breastfeeding. These questionnaires also query episodes of wheezing, and medication (over-the-counter, physician-provided samples, prescription) use and its indication.

 

DATA ANALYSIS

Measures of community diversity, richness and composition will be calculated for comparison against exposure and outcome variables. A variety of measures will be investigated. For example, the Shannon index, a measure of the difficulty in predicting the identity of the next analyzed clone, is positively correlated with diversity and evenness. Since it gives greater weighting to rare and intermediate abundance OTUs, it is more sensitive to changes in the abundance of rare groups. By contrast, the Simpson index is more heavily weighted towards dominant OTUs. It is not uncommon for both indices to be used at the same time and we intend to do the same. In addition, we will use principal components analysis (PCA) to investigate the relationship between community structure, subject exposure and outcome. Bivariate tests will be conducted to report unadjusted differences in microbiota (diversity and presence of OTU), according to overall antibiotic use and of individual antibiotic classes. Separate analyses will be conducted according to the timing of fecal sampling as follows: antibiotic use before 3 months, antibiotic use between 3 months and one year, and any antibiotic use over the course of the first year.

 

Antibiotic use will be classified as recent (≤7 days) or past administration. Microbiota diversity and presence of OTU in meconium will be evaluated against maternal antibiotic use during pregnancy. Subsequent to the large sample size it is anticipated that the Shannon and Simpson indices will be normally distributed, which will allow pair-wise comparisons for antibiotic use to be conducted with the t-test. In the event that data are skewed, medians will be compared with the Mann-Whitney U-test. Differences in percent colonization with individual OTUs will be determined with the Chi square test. To analyze the effects of antibiotic use following adjustment for other environmental factors (e.g. mode of delivery, type of diet, hospitalization after birth, maternal antibiotic use), two multivariate approaches will be pursued to accommodate the different format of the outcome variables. Firstly, linear regression will examine the effects of antibiotic use on microbiota diversity with the other factors included in models when they meet statistical significance. Second, logistic regression will be examined to determine the risk (odds ratio) of specific OTU colonization from antibiotic use, adjusted for other factors. The interaction between antibiotic use and mode of delivery, and antibiotic use and type of diet will also be tested. Both types of analyses will be conducted at the 3 month and one year time period; two models will be tested for the latter to assess the immediate and cumulative effects of antibiotics. To limit the chance of a Type II error from multiple testing, the alpha value will be set at 0.01 (2-sided) instead of the customary value of 0.05. Statistical analyses will be performed using SAS 9.1.

 

 

 

 

Development of asthma is directly related to lung function in early childhood and the impact of viral infections on the lung.

Longitudinal studies have demonstrated that decreased lung function in childhood persists into adulthood. Furthermore, poor lung function is associated with all-cause mortality, cardiovascular disease and stroke. It is possible that poor lung function in childhood determines morbidity and mortality in adulthood. Several studies have shown decreased lung function in wheezy infants, but it is unclear whether this decline persists or distinguishes between asthma phenotypes. These inconsistent findings can be partially explained by the sensitivity of methods used to measure lung function, inadequate sample sizes to detect differences, or the precise timing of infectious exposures. The timing as well as the type of respiratory viral infection may have differential effects on lung function and development of asthma. Infection with Respiratory Syncytial Virus (RSV) at a young age is a key risk factor for severe wheezing illness while rhinovirus (RV) is a pathogen in children over 2 years of age. RV is associated with increased peripheral eosinophilia and atopy implicating it as a harbinger of asthma, whereas RSV is associated with non-atopic wheezing in infants. These studies suggest there are critical windows for infections to permanently damage growing airways.

 

Infant lung function can be easily and precisely measured at birth and in the first year of life, increasing the potential for early detection of alterations in lung health during a window of opportunity for implementing interventions in at risk children.

 

Specific testable hypothesis related to INFANT LUNG FUNCTION and INFECTION.

  1. Lung clearance index (LCI) is an early sensitive marker of airway obstruction, and in infancy tracks with preschool lung function. LCI abnormalities in preschool precede abnormalities in spirometry at school age. Analyses using data from early infancy are in progress now.
  2. Maternal exposures to indoor and outdoor air pollutants, such as tobacco smoke and traffic-related air pollution, contribute to differences in pulmonary function (TPTEF/TE) measured at birth. Substantial data already collected, which can be linked with maternal environmental exposures during pregnancy.
  3. Oxidizing exposures at 3 months of age are associated with reduced lung function and increased exhaled nitric oxide at that time. Sample size – 100 with exhaled-NO measurements. Timeline – analyses 2013.
  4. Oxidizing exposures at 3 months of age are associated with reduced lung function and increased exhaled nitric oxide at 1 year of age. Sample size – 100 with exhaled-NO measurements. Timeline – analyses 2013.
  5. Lung function (LCI and spirometry) differs between infants with allergy (positive skin tests +/eczema) and healthy infants.
  6. Amongst infants and children with wheezing history; those with documented allergy will have lower lung function than those with no allergy.
  7. Incidence of viral infections amongst the general population is lower than that seen in a high risk cohort (CHILD vs. COAST)
  8. Exposures that contribute to inflammation (either systemic or respiratory) exacerbate the effects of viral or other infections on infant lung function.
  9. Severity and frequency of early life viral infections are associated with subsequent wheezing and asthma diagnosis. Sample size – the viral sub-cohort (Toronto) is approximately 800 infants. Timeline – information on viral infections and wheezing will be collected over the first few years of life. Asthma diagnosis will be made at age 5 years. These outcomes are all funded within the existing CHILD infrastructure.

 

 

The novelty of the CHILD study:

  1. Tracking of lung function in early life with repeated measures in the first five years of life with important covariates such as immune function, viral infections and exposure assessments.
  2. Characterizing lung growth in early life and determining how this relates to symptoms and phenotypes. This will help better distinguish those children at greatest risk of developing chronic disease, not only asthma in early life but other chronic lung diseases later in life.
  3. Multi-level analysis, which looks at timing of events in early life and how these relate to lung function trajectories and asthma at age 5. This is a powerful way to determine the critical periods for the development of asthma.

CHILD will be the largest longitudinal study following infants into school age with objective lung function and laboratory viral assessments. These detailed studies are undertaken in one centre (The Hospital for Sick Children, Toronto). The intent will be able to understand the role that timing of viral infection plays in the development of asthma and lung growth.

This large collection of data in healthy children also offers a tremendous opportunity to define normal lung growth and development, which will improve how infant pulmonary function data are interpreted worldwide. The normative infant data already generated have been instrumental in developing international guidelines for the interpretation of lung function.

These data have already changed fundamental concepts such as the assumption that lung function is stable in infancy. Never before has infant lung function been assessed longitudinally in such a large multi-ethnic population. It is anticipate that this study will change our understanding of the meaning of lung function measures for future lung health. Normative data from CHILD can help to inform the optimal timing of future intervention strategies designed to decrease the incidence of childhood respiratory viral infections and their impact on lung function decline.

INFANT LUNG FUNCTION and INFECTION Power and Sample Size:

With a sample size of 200 in the pulmonary function subcohort, we will have 80% power to detect an increased risk of persistent wheezing at age 5, from 20% in subjects with normal pulmonary function testing (PFT) in infancy (e.g. FEV0.5), to 33% (i.e. RR=2) when the FEV0.5 is reduced by one standard deviation below the mean, while adjusting for maternal smoking, viral infection and gender. If the probability of persistent wheezing4 is 15%, then an RR of 2.3 can be detected with 85% power. Considering a range of probabilities of persistent wheezing at the mean/normal level from 15% to 30%, Figure A illustrates that our sample size for this sub-cohort will give over 80% powers to detect RRs greater than 2.

INFANT LUNG FUNCTION Methods

The CHILD study is uniquely designed to address many of the unanswered questions regarding lung function in infancy and development of asthma. In the Toronto sub-cohort, infant pulmonary function will be measured in at least half of the study population using state of the art methods.

 

Raised volume rapid thoraco-abdominal compression technique has been shown to be more sensitive at detecting lower airway obstruction than older methods used in earlier studies. Furthermore, novel methods such as multiple breath washout, which is a more sensitive measure of small airway obstruction are being utilized. Inflammation measured using tidal breathing exhaled nitric oxide is assessed at birth and at all infant lung function assessments. In Toronto, infant lung function is assessed on three occasions 3 months, 1 year and 2 years. We will then measure lung function annually at 3, 4 and 5 years using spirometry and multiple breath washout.

 

Preschool spirometry has been proposed for the entire cohort at ages 3 and 5 years. Thus, trajectories established in the Toronto cohort may be applied to the entire cohort and thus we will be able to maximize on the data collected nationally to look at factors involved in the development of asthma.

 

Furthermore in the CHILD Toronto sub-cohort we will be taking viral swabs not only at screening visits at 3 months and 12 months but also at times of acute viral lower respiratory infection in the first year of life (See Appendix D). Our study is the first to prospectively study the role of RSV and other viruses in the development of wheezy illness in a general population cohort by looking at immune, lung function and airway inflammatory markers not only post infection but, most importantly, pre-infection. See a video of the Infant Pulmonary Function Lung Assessment at SickKids

 

RESPIRATORY TRACT INFECTIONS Methods

Respiratory tract infections are being assessed in the full cohort and in more detail in the Toronto sub-cohort.
Direct viral detection: A nasal swab sample, collected in all children at 3 months and 1 year of age, is separated into 3 aliquots and frozen at -80oC. Subsequently, with separate funding, nucleic acids will be extracted from the aliquots 1 and 2 using the bioMérieux MiniMag extractor (bioMérieux, Marcyl'Etoile, France). The ID-TagTM respiratory viral panel (RVP) (Tm Bioscience [Luminex] Corporation, Toronto, Ontario) bead-based microarray method will be used to detect the presence of common respiratory viruses (influenza A and B, parainfluenza types 1-4, RSV, enterovirus/rhinovirus, coronaviruses, metapneumovirus and adenovirus) in aliquot 1. For samples which are negative by the RVP, aliquot 2 will be tested on the Virochip microarray which tests for all known viruses simultaneously. The third aliquot will be used for virus-specific PCR tests to confirm the results of the RVP or Virochip.

 

In the Toronto sub-cohort, viral infections are being monitored and captured by two methods. A respiratory symptom diary is given to the mother at birth, and information transferred to study records at each follow-up visit. Exposure to other children is assessed through bi-annual time/activity logs that include daycare questions, age when the child first started daycare, and hours/days spent at the daycare. Other infections (urinary tract, gastrointestinal, and otitis media) are obtained by questionnaire every 6 months.

 

ACUTE VIRAL INFECTIONS Methods

Acute infections in the first year of life are reported by parents to the Toronto CHILD study team via a respiratory infection call hotline. A symptom assessment is administered via standardized questionnaire (URECA questionnaire 6) by a nurse. Lower respiratory infections of sufficient severity as assessed by questionnaire will result in a clinical assessment of the child by a trained nurse and administration of an acute viral swab.

 

 

 

Early childhood socioeconomic status (SES), through life stressors, material deprivation, and inadequate living conditions, interacts with the physical environment and genotype to influence the development of asthma.

 

SES has a profound influence on physical health over the lifespan. Children of parents with lower SES have more severe asthma, higher rates of hospitalizations and more days in bed due to asthma. Stress may also contribute to the development and expression of asthma. Parental stress forecasts the onset of wheezing in infancy. Stress may be especially detrimental when it involves interpersonal difficulties between family members.

 

To understand how psychosocial factors might affect asthma development, we will draw upon recent discoveries establishing functional connections between the nervous, endocrine, and immune systems. These connections enable the central nervous system to "transduce" the social environment into hormonal signals, propagated systemically by the autonomic nervous system and the hypothalamicpituitary- adrenocortical axis, and other endocrine circuits.

 

Most classes of leukocytes have receptors for the hormonal products of these circuits, enabling more severe, chronic stressors to elicit changes in multiple aspects of adaptive immunity, including blunted antibody responses to vaccination, heightened susceptibility to viral infections, delayed healing and disrupted cross-talk between Th1 and Th2 cytokines.

 

Stress is not globally immunosuppressive, however, as it induces a mild, chronic inflammation by accentuating the activity of a key pro-inflammatory transcription factor, nuclear factorkappa B (NF-KB).These discoveries provide a basis for unraveling the mechanistic underpinnings of "mind-body" effects. The CHILD Study, with its broad investigative and population base and detailed psychosocial assessments, provides the ideal opportunity to examine these relationships.

 

Specific testable hypotheses related to PSYCHOSOCIAL ENVIRONMENT:

  1. Children who are reared in low-SES and/or high-stress families will display intermediate phenotypes (as manifest in epigenetic, transcriptional, and immunologic profiles) that predispose them to subsequent asthma. Work on this hypothesis is in process. As part of the REEGLE and Microbiome projects, analysis of cord-blood and one-year samples has begun on a subset of ~250 offspring. In the coming months these analyses will yield profiles of DNA methylation and ex vivo cytokine production. At that point, we will be positioned to examine:
    • how the in utero psychosocial environment relates to intermediate phenotypes at delivery and one year of life, and
    • the extent to which these influences are modulated by the psychosocial environment during infancy.
  2. Exposures in the social (low SES, high stress) and physical environment (traffic related air pollution, phthalates, and secondhand tobacco) will act synergistically to predispose children to risky intermediate phenotypes and adverse clinical outcomes. With the emerging laboratory data from REEGLE and Microbiome, we will soon be ready to evaluate the "intermediate phenotype" portion of this hypothesis, using social and physical environment data collected prenatally, and biological phenotypes at birth and one-year. The analysis of clinical outcomes (asthma, allergy) will have to wait until full ascertainment of case status.
  3. Children who develop transient wheeze, allergy and asthma are more likely to have been reared in homes with low SES (limited financial resources, low educational attainment) and/or high stress (recurring parental conflict, familial instability, financial/employment difficulties.) Testing this hypothesis will have to await full ascertainment of clinical status at 3 years and 5 years.

The novelty of the CHILD study:

CHILD includes a detailed assessment of the pre- and post-natal psychosocial environment. The scale and scope of these assessments make CHILD unique among birth cohorts in the allergy/asthma field.

 

When these data are linked to CHILD's epigenetic, transcriptional, microbiologic, immunologic, and phenotypic outcomes, they will provide an unprecedented opportunity to unravel the mechanisms through which psychosocial characteristics "get under the skin" to affect risks for allergy/asthma. The psychosocial assessment focuses on the family's socioeconomic status, the level of stress present in the home, and the quality of the parent-child relationship.

 

CHILD's rich psychosocial data will be related to the newborn's gut microbiota, innate and adaptive immune responses, and cord-blood methylation/transcription profiles, i.e. epigenetic factors.

 

As the child grows, these risk factors will also be related to trajectories of immune response and pulmonary function, with the idea that high levels of stress will confer risk for a phenotype characterized by excessive cytokine responses to allergens and decreased growth in pulmonary flow measurements. Clinically, these changes will be related to sensitization and the development of wheeze. SES can be subjective or represented by financial resources or parental education.

SOCIO-ECONOMIC STATUS (SES) PSYCHOSOCIAL - Sample Size / Power

In the Stress sub-cohort, a sample size of 400 will provide 80% power to detect an increase in risk of developing atopy at age 5 from 20% to 29%, when the family stress level is increased by one standard deviation above the mean (i.e. high chronic stress). This corresponds to a relative risk (RR) of 1.64. If RR=1.70 is considered, the proposed sample size will achieve 85% power at a significance level of 5%.

 

If atopy at age 5 is less prevalent (e.g. 15%), then with this sub-cohort sample size, RRs of 1.9 and 2.0 can be detected with powers of 90% and over 95%, respectively. If the prevalence of atopy at age 5 is assumed to be 25%, these RRs can be detected with even greater powers, as shown in the figure below. Based on the observed retention rate (95%) thus far, we expect the attrition rate to be lower than originally estimated, leading to greater power for hypothesis testing.

Figure B Psychosocial

SOCIO-ECONOMIC STATUS (SES) Methods

 

Each method of characterizing SES will be evaluated independently and jointly within the analyses.

 

Subjective SES:

Parents are asked to place their household (relative to their community) on a ladder using the MacArthur Scale of Subjective Social Status.

 

Resource-based SES:

Parents are asked about family assets (e.g. income, savings, housing status (rent/own), number of bedrooms, and number of cars). Responses will be standardized (z-scored for the sample) and the z-scores summed to create one composite asset score.

 

Prestige-based SES:

Years of education and highest educational degree attained for each parent is recorded. Parental job characteristics are also recorded and will be used to index occupational prestige using standardized Canadian coding systems. Psychological stress will be assessed along a number of dimensions, both prenatally (at 18 and 36 weeks gestational age (GA) and annually from the time of the index child's birth:

Chronic stress in multiple life domains will be assessed including; the nuclear family, the broader family/social network, in work and/or academic settings, or from health problems in self/family. Acute life events in family, social network, work life, health domains will be assessed. Perceived stress, reflecting the extent to which life is viewed as unmanageable, and unpredictable, will also be assessed. Multiple aspects of maternal mood will be measured, including symptoms of depression, and positive and negative mood states. Factors that can protect against stress, or accentuate its impact, will also be measured. These factors include social support, social capital, marital quality, and social integration.

 

The following standardized instruments are currently being employed in the CHILD Study. Post-natal questionnaire sets are identical to the 36 week prenatal set with the addition of the Parenting Questionnaire:

CPre-natal: 18-34 Weeks GA

  1. Profile of Mood States (POMS)
  2. Perceived Stress Scales (PSS)
  3. Centre for Epidemiological Studies - Depression Scale (CES-D)

Pre-natal Assessment 36 (± 4 ) Weeks GA

  1. Chronic Stressor Evaluation
  2. Recent Life Events (RLE)
  3. Profile of Mood States (POMS)
  4. Perceived Stress Scales (PSS)
  5. Centre for Epidemiological Studies - Depression Scale (CES-D)
  6. Dyadic Adjustment Scale (DAS)
  7. Interpersonal Support Evaluation List (ISEL)
  8. Socio Economic Status Index (SES)
  9. MacArthur Subjective Community Standing Scale
  10. MacArthur Subjective Socioeconomic Standing Scale

Post-natal: 1 year and subsequently

  1. Chronic Stressor Evaluation
  2. Recent Life Events (RLE)
  3. Parenting Questionnaires (PQ)
  4. Profile of Mood States (POMS)
  5. Perceived Stress Scales (PSS)
  6. Centre for Epidemiological Studies - Depression Scale (CES-D)
  7. Dyadic Adjustment Scale (DAS)
  8. Interpersonal Support Evaluation List (ISEL)
  9. Socio Economic Status Index (SES)
  10. MacArthur Subjective Community Standing Scale
  11. MacArthur Subjective Socioeconomic Standing Scale
 A link to questionnaires developed for the CHILD Project

The impact of the physical environment on the development of childhood allergy and asthma, through modulation of the developing immune system, can be explained by exposures that collectively contribute to chronic inflammation via oxidative stress pathways.

 

Specific testable hypotheses related to the physical environment: INDOOR AIR:

Multiple exposures have been associated with wheeze and asthma in early life (environmental tobacco smoke, oxides of nitrogen, allergens particularly dust mite, mould and cat, and endotoxin).


  1. Endotoxin has been linked with increased inflammatory responses during viral infections, but also with decreasing asthma in specific farming environments. Indoor airborne dust is associated with increased respiratory symptoms and irritation to mucous membranes, in part due to volatile organic compounds (VOCs) on dust particles.
  2. Indoor chemicals have been associated with the development of asthma. Innate immune inflammatory response can be activated by common indoor exposures, including trichloramines, aldehydes (e.g., formaldehyde) and other VOCs found in home furnishings and cleaning products.
  3. Phthalates, chemicals with many applications from building materials to personal care products, are implicated in asthma and allergy. Persistent organic chemicals such as perfluorinated substances, although primarily a risk in relation to reproductive and developmental effects, can also elicit an immune response.

 

Specific testable hypotheses related to the physical environment: OUTDOOR AIR:

Ambient air pollutants (particulate matter, nitrogen dioxide, sulphur dioxide, ozone and carbon monoxide), VOCs, semi-VOCs (SVOCs), and aeroallergens can cause significant respiratory morbidity. Even at the low ambient levels in British Columbia, an association between these pollutants and development of asthma has been detected. However, Traffic- Related Air Pollutants (TRAP) were associated with the highest risks. TRAP asthma exacerbations occur through multiple mechanisms including oxidative stress. More importantly, evidence for incident asthma due to TRAP is increasing. The effect of TRAP on airways disease is likely modified by genetics and co-exposures such as allergens. CHILD preliminary data (link to traffic data table) show substantial variability in TRAP across centres as judged by Land-Use-Regression derived NO2 estimates at CHILD study participants homes.

 

Indoor and outdoor exposures, including TRAP, can be chronic or transient. We do not understand their relative and/or combined importance in the development of non-atopic asthma, and the mechanisms involved. Given that their individual risks are small even in high risk groups, it is plausible that multiple exposures, indoor and outdoor, act together via oxidative-stress pathways that involve airway epithelium and inflammatory cells. Furthermore, these exposures in combination with other stressors (e.g., infection, allergen exposure, stress), pose the greatest risk during key stages in immune system development, particularly when immune balance is compromised. These interactions have not been comprehensively explored in any birth cohort to date. The analytical approaches to CHILD data will enable examination of these multiple risk factors with a level of detail not previously available. Risk strata based upon exposure can be assembled prior to the emergence of disease and then tested against disease outcome at age 5. This will determine the most robust predictive exposure models and indicate how much oxidative-stress inducing exposures contribute to the development of asthma in childhood. For details of methods see Table 8 and Appendix G.

 

Specific testable hypotheses related to ENVIRONMENTAL EXPOSURES :

  1. Pregnancy results in enhanced systemic innate (TLR, RLR, NLR) immune capacity and exposures related to the physical environment that can induce systemic inflammation (e.g., environmental tobacco smoke, air pollution) are associated with differences in the magnitude and nature of the enhancement. A longitudinal study, (n=400 women), which is already in progress, will use CHILD blood in Winnipeg with data from the 18 week environment questionnaire.
  2. Mother's exposures related to her physical environment that can induce systemic inflammation (e.g., environmental tobacco smoke, air pollution) are linked to maternal innate immune capacity and its return to the non-pregnant state of homeostasis. Study in progress using CHILD blood in Winnipeg (n=400) with data from the 18 week environment questionnaire.
  3. Organic compounds indicative of motor vehicle emissions are present in house dust and provide a surrogate of exposure that incorporates infiltration of outdoor air and local vehicle behaviours not captured by land-use regression.
  4. Time in traffic and other time activity related behaviours can be used to refine traffic related air pollutant (TRAP) exposure estimates and to identify mothers and babies experiencing the greatest exposures to TRAP. This involves correlation of a Land Use regression of outdoor NO2 from latitude and longitude, time activity questionnaire data from pregnancy and early infancy. Timeline for results = 2013
  5. Exposure to endotoxin and beta-glucan in the home varies by season and among cities. These analyses will use data obtained through CMP funding and CHILD questionnaires.Timeline for results = 2013-14.
  6. Increased exposure to perfluorinated compounds is associated with an increased risk of wheezing episodes by age 1 year. Timeline for results = 2013.
  7. Exposures to oxidizing agents in the prenatal period influence birth-weight. Timeline for results = 2013.
  8. Exposure to allergens in the home varies by season and among cities. Timeline for results = 2014.
  9. Perinatal exposures to oxidizing agents are reflected in elevations in biomarkers of inflammation in cord blood, specifically IL-6. Sample size ~500. Timeline: analyses in 2013-14 provided funding obtained.
  10. Cooking with gas stoves is a risk for wheeze in early childhood. Sample size – 1000 children at age 3 years. Timeline: analysis can be conducted in 2013-14.
  11. Immigrant children, including first and second generation immigrants, have low prevalence of wheeze at 1 year of age compared to the native children in developed countries. Sample size: entire cohort. Timeline 2013-14.
  12. Cooking with gas stoves is a risk for persistent childhood asthma. Sample size – all 300 with physician diagnosed asthma at age 5, and 2:1 controls = 900 children. Timeline: outcomes will not be available until age 5 years.
  13. Environmental exposure to pesticides increases the risk of early childhood wheeze. Sample size – 1000 children at age 3 years. Timeline: 2013-14.
  14. Inflammatory markers evident in early development (birth, 3 months) predict the development of atopy and wheeze at 1 year. Sample size ~ 1000. Timeline 2013-14.
  15. Exposures to oxidizing agents in the prenatal period and during the first 3 months of life influence the development of atopy and wheeze at 1 year. Sample size ~ 1000. Timeline 2013-14.
  16. Second- and third-hand smoke exposures vary among infants and can be assessed by questionnaire. This will involve correlating questionnaire inflammation during pregnancy and early childhood with urine cotinine measurements in infants. Sample size ~ 500. Timeline 2013.
  17. Infant phthalate exposures (assessed in house dust) vary be season and are greater in cities with colder climates; these differences are reflected in urine metabolites. Sample size ~ 1000. Timeline 2013.
  18. Phthalate exposure measured by urine metabolites and dust levels of the parent compounds can be predicted by exposure questionnaire data.
  19. Exposure to Traffic-Related Air Pollution (TRAP), phthalates, cleaning products and tobacco are influenced by socioeconomic status.
  20. Children's ranking of distribution of phthalate metabolite levels in urine will be similar between analyses at 3 months and 1 year indicating that a urine sample at a single time point is indicative of chronic exposure due to stable behaviors by mom, baby and family and persistence of phthalates in homes.
  21. Phthalate metabolite levels in urine will increase from 3 months to 1 year related to increased exposure resulting from changing behaviours e.g. child mobility and diet.

 

The following links take people to tables highlighting the finding to date.

link to phthalates data table | link to traffic polution data table | link to burden of dust data table | link to housing demographics

EXPOSURES - Power / Sample Size

 

The timeline for results and the sample sizes for the specific exposures sub-hypotheses are presented in the main text above this sub-section. For further clarification, phthalate BBzP has been implicated in the induction of allergy and asthma. Assuming an exposure level of 0.157 (0.139-0.178) mg/g dust [geometric mean and 95%CI] for the unaffected individuals, and based on the concentrations reported by Bornehag et al.5, the table below shows the powers to detect 5, 10, 15 and 20% differences in exposure levels, calculated using the t-test, between affected and unaffected individuals for outcomes that occur in 5, 10, 20, 30 and 40% of children. The powers shown in brackets are based on the Mann-Whitney U-test.

 

Outcome \ Difference 5% 10% 15% 20%
5% 28 (27) 79 (77) 98 (98) 99 (99)
10% 50 (48) 97 (96) 99 (99) 100 (100)
20% 79 (77) 99 (99) 100 (100) 100 (100)
30% 92 (91) 100 (100) 100 (100) 100 (100)
40% 97 (97) 100 (100) 100 (100) 100 (100)

 

As an example, when considering the indoor air physical environment, our sample size will give rise to greater than 80% power to a difference of 0.02 in endotoxin concentration within dust (EU/g) between group means-asthmatic vs. non-asthmatic when using a two-sided Mann- Whitney test and assuming that the non-asthmatic geometric mean is 4.6 EU/g. To account for possible non-environmental effects, such as sex, age and sib-ship size, we will also use matched pairs (asthmatic vs. non-asthmatic) to examine the relationship between endotoxin and the development of asthma6. With 300 matched pairs, we will have more than 80% power to detect an OR of 1.7 per SD in endotoxin, using conditional logistic regression.

EXPOSURES Methods

Indoor and outdoor environmental exposures are assessed by multiple methods, including questionnaires, home inspections, dust sampling for multiple allergens and pollutants and outdoor exposure modeling. The timeline for results and the sample sizes for the specific exposures sub-hypotheses are presented in the main text above this sub-section.A brief overview of each collection method follows:

 

1. Questionnaires:

A comprehensive baseline (during pregnancy) and comprehensive updates at 3 months, 1 year, 3 years and 5 years of age are administered regarding address, housing structure, function, condition, maintenance history and cleaning habits, presence and use of attached garage, renovations, source and extent of dampness indicators, mould growth, new furnishings, appliance and household cleaner emissions, presence and type of air conditioning and cleaning, conditions of use, etc. Questions related to the child's and families' time activity include time spent in different rooms in the house and indoors vs. outdoors, time in transit, mode of transport, frequency/duration of visits to daycare, indoor pools, other microenvironments, exposure to smoke, etc. Shorter followup questionnaires are administered at 6 and 18 months, 2, 2.5 and 4 years of age and focus mostly on time activity of the child and major changes such as home renovations. A list of questionnaires developed for the CHILD project.

 

2. Environmental Tobacco Smoke:

Environmental tobacco smoke exposure is assessed on the comprehensive prenatal and postnatal questionnaires with validation in a subset by cord blood and 3 month urinary cotinine measurements.

 

3. Home Assessment:

When the infant is approximately 3 months old, RAs trained by Canada Mortgage and Housing Corporation (CMHC) and lead CHILD environment investigators visit the home to measure geographical coordinates (GPS), evaluate the structure, function and exposure sources within the home with specific emphasis on the evaluation of the house in terms of cleanliness and cleanability, furnishings, ventilation (including presence of air-conditioning and type of heating/cooling), humidification, microbial and chemical contaminant burdens, and basement conditions. A link to a description of the Home Assessment manual developed for the CHILD study.

 

4. Dust Sample Collection:

During the home assessment, a dust sample is collected from a 2 m2 area of flooring using a standardized consumer model vacuum cleaner fitted with a specially designed dust collection device. The measured area is enlarged as necessary to collect a minimum of 1-2 g fine dust in a standardized manner for use in allergen and endotoxin testing, as well as markers of indoor chemicals (e.g., phthalates) and infiltration of outdoor traffic particles. The dust sample is collected from the child's bedroom (bed and floor) and the living space where the child spends the most time. Total dust weight and fine dust weight is determined and multiple aliquots are prepared and stored at -80ºC. At present, analytical protocols use less than 100 mg of dust and thus a considerable amount is archived for future considerations.

 

5. Dust sample analysis for allergens:

One aliquot of fine dust (at least 50 mg) is extracted at room temperature for 2 hr using a standard protocol. After analysis for endoxin and B glucan the remainder is used for allergen analysis with standard ELISA protocols for Der p 1, Der f 1, Mite group 2, Fel d 1, Can f 1, Mus m 1, Rat n 1, Bla g 2, Alt a 1, Bet v 1. Dust analysis for semi-volatile organics: Samples are analyzed similar to what has been reported previously by Langer et al. (2010). Briefly 10 mg of sieved dust is extracted in solvent using Accelerated Solvent Extraction (ASE), concentrated and then subjected to GCMS to quantify phthalates, hopanes (Brook et al., 2007) and PAHs in ng/g of fine dust.

 

6. Outdoor Exposure Modeling:

Family residence history and time-activity information is collected through the questionnaires. Outdoor air exposure modeling will be based upon available monitoring data and city-specific land-use regression (LUR) models focused mainly on NO2 (Kanaroglou et al, 2005; Henderson et al., 2005). More advanced data fusion/assimilation methodologies (Brook et al., 2010) using all available data are constantly evolving within the CHILD/AllerGen network and Canadian exposure research community. This includes government monitoring data, LUR output and air quality model output and satellite-based observations. In time series fashion, these data will be linked to the geographic information collected on the family and combined with indoor exposure information. These assessments will include models that take into account outdoor-indoor exchange of pollutants given the amount of time the children spend indoors.

 

7. Urine Sample Collection:

Urine is collected at the 3 month home visit and at the 1, 3 and 5 year clinics. At the beginning of the 3 month and 1 year visits, a tegaderm liner is placed in the child's diaper followed by the placement of cotton pads to absorb the urine. At the end of the visit (approximately 2 hours) the urine is collected by placing the cotton pads in a large syringe and squeezing out 0.5 ml of urine into 6 pre-labelled cryovials. The last drop is placed on a refractometer and the specific gravity recorded. Samples are kept cold and frozen within 8 hours. At 3 years, urine is collected in a similar manner for those children not yet toilet trained. For those who are toilet trained, a potty, with an appropriate insert, will be provided in the clinic. A volume of 1.5 ml of urine is stored in each of 6 cryovials.

 

8. Breast Milk Sample Collection:

Breast milk is collected from mothers using a collection jar provided with instructions prior to the 3 month home visit. Hand expressed breast milk is collected from 1 day prior to the visit to include at least 2 separate feedings with milk collected before and after feeding. A total of 10 mls of breast is dispensed by study staff into 6 x 1.5 ml aliquots and frozen within 8 hours.

 

9. Stool Sample Collection:

Meconium is collected at birth, by a nurse who transfers the meconium from the baby's diaper into a disposable container. Stool is collected by mothers who provide a soiled diaper at the 3 month and 1 year visits. The meconium/stool specimen collected is divided equally into 4 preweighed cryovials and frozen within 8 hours.

Aliquots are shipped weekly to the the Clinical Research and Clinical Trials Laboratory (CRCTL) for long term liquid nitrogen storage.

Specific gene-environment interactions predict the development of allergic and asthmatic phenotypes.

 

Asthma and other allergic diseases are the result of complex interactions between multiple environmental factors and genetic variants that confer susceptibility. Large-scale genome-wide association studies (GWAS) searching for susceptibility loci for asthma and related traits have identified several genes with robust associations (Figure 3). GWAS data from ongoing meta-analyses of international consortia will likely uncover more novel susceptibility genes. Intermediate phenotypes related to allergic diseases, such as peripheral blood eosinophil counts have also been investigated by GWAS and detected associations not seen with studies of asthma. However, a substantial portion of the heritability of these phenotypes remains unexplained and is likely due to interactions between environmental factors, genes, age, sex and epigenetic effects.

 

Over 25 consistently replicated genes and over 15 separate environmental exposures have been associated with the development of atopy and asthma, resulting in large numbers of gene-environment combinations to test. However environmental exposures may impact asthma differently at different time points in life, and the relevant risk factors may change over time, further adding to the complexity. Simultaneous prospective longitudinal measurements of genetic and environmental risk factors for atopy and asthma are necessary to develop knowledge translation strategies to prevent these diseases.

 

The most frequently reported gene-environment interaction studied involves the CD14 -159 C to T variant (rs2569190) in the promoter region of the gene and exposures to endotoxins. Over 200 studies have examined the association between CD14 and asthma related phenotypes, but results have been inconsistent. This may be explained by a complex dose-dependent environmental interaction with endotoxin exposure in early life, whereby the T allele can either confer protection or increase risk depending on the level of exposure. These interactions may obscure or confound results from studies that include only main genetic or environmental effects and underscores the need to include interactions in statistical models for complex disease phenotypes such as asthma. In addition to the environmental factors described above, genetic factors related to host susceptibility and response to viruses will be tested for association with viral infection and measures of the innate immune response. Genetic factors will also be associated with specific aspects of the metagenomic data being gathered in CHILD, e.g. with respect to susceptibility to infection with H. pylori. Finally, geneticpsychosocial factor interactions will be sought to predict clinical outcomes in CHILD.

 

Specific Testable Hypotheses Related to Genetics:

 

Genetic variants that were robustly associated with asthma in genome-wide case-control studies are responsible for a high proportion of the population attributable risk for asthma / allergy in the general population.

  • The GWAS reported to date have identified numerous, genome-wide significant genetic associations with asthma but these have not been tested in a prospective cohort from the general population. CHILD will enable this critical analysis, minimizing the loss of power due to correction for multiple comparisons that is incurred in a GWAS, and applying results to many other clinical outcomes of interest (e.g. food allergy). Sample size – 3000 children with DNA samples available. There are multiple timelines: analyses of intermediate outcomes such as atopy will be performed at ages 1, 3 and 5 years. Analysis of probable asthma will be performed at 5 years of age.
  • GWAS have also been performed for atopic dermatitis in European and Asian populations and we will test genome-wide significant associations in the CHILD cohort. The specific genes are shown in Table 2 (end of document) and these will be tested for association with atopic dermatitis at ages 1, 3 and 5 years.
  • A panel of 3000 genetic variants consisting of those listed in Tables 1 and 2, plus other polymorphisms in the same genes / chromosomal regions that showed independent associations with asthma, will be genotyped in the children of the CHILD cohort. We will use an iSelect HD Custom Genotyping BeadChip (Illumina) and funding for this ($220,000) will be sought from CIHR. Potential alternative sources of funding for a smaller panel include an AllerGen NCE Strategic Initiative, and provincial lung associations.

Genetic variants associated with asthma in GWAS and candidate gene studies interact with key environmental risk factors for asthma / allergy. A limited number of studies have tested interactions of genes with environmental factors in the pathogenesis of asthma / allergy. The depth of environmental assessment in the CHILD study makes it uniquely placed to investigate gene-by-environment interactions.


  • Polymorphisms interact with RSV infections to influence infant lung function. We will genotype the CHILD infants in whom lung function measurement have been made for polymorphisms associated with susceptibility to RSV infection (IL8, IL19, IL20, IL13, MBL2, IFNG, and RANTES) as well as variants in the RSV receptor (NCL). Timeline: 2013-14, depending on availability of lung function and RSV data.
  • Polymorphisms interact with the environmental variables measured in CHILD to influence innate immune development. We will assess the impact of genetic variants on innate immune function of cord blood samples available in the CHILD cohort. Polymorphisms will be examined in the PRR genes and other targets downstream in the signal transduction pathways. Timelines depend on availability of innate immune readouts.
  • Polymorphisms interact with antibiotic exposures in utero and during childhood to increased risk for wheeze or asthma later in childhood. A stronger antibiotic effect was observed in infants without a family history of asthma suggesting that genetic factors interact with antibiotic exposures. We will genotype the CHILD infants for polymorphisms in key innate immunity genes involved in regulation of the intestinal microbiota (e.g. TLRs, MyD88). Timeline: this requires primary outcome data at 5 years of age.
  • Polymorphisms shown to influence adult lung function also modulate infant airflow obstruction. Several GWAS have examined the genetic influence on lung function in general population samples but most of these studies utilized cohorts of adults. Interestingly, some of the identified genes are thought to play a role in lung development (e.g. HHIP, GPR126, RARB) and in one study 69% of the associations with lung function in adults were consistent in children (7-9 years of age). These data suggest that there are important genetic effects on lung development and we hypothesize that stronger effects will be observed at earlier ages, as the influence of environmental factors will be less. Timeline 2013-14, depending on availability of lung function data.
  • Genetic polymorphisms in TSLP, the TSLP receptor and TLR genes are related to CD34+ hemopoietic progenitor eosinophil / basophil 'lineage priming' at birth. We will examine the relationship of polymorphisms in TSLP, IL-7R (subunit of the TSLP receptor) and TLR genes (TLRs 1-10) to cord blood CD34+ progenitor eosinophil / basophil phenotype and functional differentiation responsiveness (i.e. eosinophil / basophil 'lineage priming') in CHILD subjects. Sample size = 400 cord blood samples. Funding from CIHR via operating grant to Dr. Denburg.
  • Genetic polymorphisms influence the DNA methylation patterns in the human genome relevant to the pathogenesis of asthma / allergy. Several studies have shown that the DNA methylation state at certain loci in the genome is highly heritable and several specific genes have been implicated. We will collaborate with the investigators on the REEGLE project to determine the influence of genetic polymorphisms on the DNA methylation patterns found to be associated with asthma / allergy in the CHILD samples. Timeline: 2014-16.
  •  

The novelty of the CHILD study:

CHILD will provide a unique opportunity to test the impact of gene-environment interactions on the development of allergic disease phenotypes. We will test specific genetic factors robustly identified (p values <10-7) in the most powerful GWAS published to date and polymorphisms that have already been shown to have interactions with environmental factors. These genetic factors will be typed in individuals who also have exquisite phenotyping in a longitudinal assessment, providing unprecedented power to detect gene-environment interactions.

GENETICS - Power / Sample Size

The table (CHILD STUDY TABLE G3: SAMPLE SIZE CALCULATION TO DETECT GENE BY ENVIRONMENT INTERACTIONS) below shows the necessary sample size at 80% power to detect various strengths of gene x environment interactions using trios, case-control analyses and case-only analyses. The examples show results for a log-additive model, with estimates for a dominant model in parentheses. The relative risks of the genotype alone and the environmental factor alone are assumed to be 1.5, and prevalences of the environmental exposure (E) equal to 5% and 10% are considered.

For gene-environment interactions, the case-only approach is the more efficient design, but it has assumptions that are difficult to check. For trio and case-control designs, we should be able to detect interactions when the prevalence of the environmental exposure is ≥10% and the interaction relative risk is ≥ 2. The dominant model always requires somewhat more subjects than the log-additive model.

 

CHILD STUDY TABLE G3: SAMPLE SIZE CALCULATION TO DETECT GENE BY ENVIRONMENT INTERACTIONS

Allele frequency
RR for interaction
# trios: E =10%
# cases & controls: E =5%
# cases & controls. E =10%
# cases in case-only design: E =5%
10%
2.0
943
2344 (2942)
1308
595
2.5
491
1261 (1608)
716
288
3.0
318
839 (1082)
484
175
4.0
181
497 (651)
296
90
20%
2.0
534
1431 (2236)
809
512
2.5
277
787 (1254)
317
248
3.0
181
535 (864)
246
151
4.0
331
205 (540)
107
77

 

The Benjamini and Hochberg method for estimating the false discovery rate (candidate genes) will be used to account for multiple testing.

GENETICS Methods

Gene-environment interactions identified in the previous literature (CHILD STUDY TABLE G1: CANDIDATE GENE STUDIES SHOWING GENE-ENVIRONMENT INTERACTIONS) and in the ongoing metaanalyses in consortia such as EVE and GABRIEL, will be examined in the CHILD participants. In addition, we will test polymorphisms in genes that have been robustly associated with asthma in published GWAS i.e. those showing overlap between the large-scale studies. For the gene-environment interactions we will test those polymorphisms in the Toll-like receptor genes previously associated with asthma and related traits (CHILD STUDY TABLE G2: POLYMORPISMS IN THE TLR GENES PREVIOUSLY ASSOCIATED WITH ASTHMA AND RELATED TRAITS).

 

Genetic and environmental factors will be associated with the outcome of physician-diagnosed asthma at 5 years of age and we expect that ~10% of the CHILD participants i.e. 500 children will have this outcome. We will also utilize wheeze as an outcome that is expected to have a prevalence of 17-20% at 5 years of age [Thomas 2010]. Additional phenotypes will be other allergic diseases such as atopic dermatitis.


Blood samples are being collected in EDTA tubes and buffy coat white blood cells are stored frozen until DNA extraction. The DNA extraction will be conducted in the laboratories of Drs. Paré and Sandford at the UBC James Hogg Research Centre at St. Paul's Hospital, Vancouver using DNeasy Blood & Tissue Kits (Qiagen). DNA samples will be quantified using Quant-iT™ PicoGreen dsDNA Reagent (Invitrogen) and normalized using a Biomek FX Liquid Handling System (Beckman).

 

Genotyping for the selected single nucleotide polymorphisms (SNPs) will be performed using an Infinium iSelect HD Custom Genotyping BeadChip (Illumina) under a service contract with the McGill University and Génome Québec Innovation Centre. SNPs will only be selected for inclusion on the chip if the minor allele frequency is greater than 5% in Caucasian populations. The allele frequencies will be checked in the CHILD population and only those SNPs with minor allele frequency >5% will be included in the analysis. Each polymorphism will be checked for agreement with Hardy-Weinberg equilibrium and excluded if p<0.001 in the unaffected controls.

 

Each clinical phenotype will be investigated using logistic regression with genotypes considered under an additive model. We will test for gene-environment interactions by including a product term in these models, with the null hypothesis that each potential risk factor (genotypes and environmental exposures) is independent. Due to the large number of comparisons we will estimate the false discovery rate i.e. the proportion of false-positive results in the set of rejected hypotheses [Benjamini and Hochberg 1995]. Gene-environment interactions will be tested in the entire CHILD population including ethnicity as a covariate and in the Caucasians only to guard against false positive associations.

 

CHILD STUDY TABLE G1: CANDIDATE GENE STUDIES SHOWING GENE-ENVIRONMENT INTERACTIONS

Gene Gene symbol Polymorphism 1 Environmental factor Outcomes Reference
ADAM metallopeptidase domain 33 ADAM33 rs528557 or rs3918396 In utero cigarette smoke exposure Bronchial hyper-responsive ness Reijmerink
B-2-adrenergic receptor ADRB2 rs1042713 (Arg16Gly) Cigarette smoke Asthma Wang
Clara cells 16 kDa secretory protein CC16 rs3741240 (A38G) East vs. West environment Rhinitis Zhang
CD14 CD14 G+1437C / rs2569190 (C-159T) Cigarette smoke Asthma severity / IgE levels Choudhry
CD14 CD14 rs2569190 (C-159T) House dust mite exposure Asthma Zambelli-Weiner
CD14 CD14 rs2569190 (C-159T) Endotoxin exposure Atopy / eczema / nonatopic wheeze Simpson
CD14 CD14 rs2569190 (C-159T) Farm exposure Nasal allergies / atopy Leynaert
CD14 CD14 rs2569190 (C-159T) / rs2569191 (T-1145C) Exposure to pets Atopy Bottema
CD14 CD14 rs2569190 (C-159T) Country living in childhood Asthma Smit
CD14 CD14 rs2569190 (C-159T) Helicobacter pylori infection Total serum IgE Virta
CD14 CD14 rs2569190 (C-159T) East vs. West environment Itchy rash Zhang
Chitotriosidase CHIT1 rs2486953 or rs4950936 or rs1417149 Mold exposure Childhood asthma exacerbations Wu
High affinity IgE FCER1B rs569108 (Glu237Gly) Day care attendance IL5 level Hoffjan
Low affinity IgE receptor FCER2 rs28364072 (T2206C) Inhaled cortico-steroid use Asthma exacerbations Tantisira
Filaggrin FLG R501X / 2282del4 Exposure to cats Eczema Bisgaard
Glutathione S-transferase M1 GSTM1 Null variant In utero smoke exposure Childhood asthma Rogers
Interleukin-1 receptor antagonist IL1RN rs2234678 Maternal smoking during pregnancy Childhood asthma Ramadas
Interleukin-4 receptor IL4RA rs1805010 (Ile75Val) / rs3024571 (Asn142Asn) Day care attendance LPS-induced IFN- Y / Atopic dermatitis Hoffjan
Interleukin-12 IL12RB1 rs425648 Breast-feeding Food sensitization Hong
Receptor B1
Interleukin-13 IL13 rs20541 (Arg144Gln) Birth order Atopy Ogbuanu
Nitric oxide synthase 3 NOS3 rs1799983 (Glu298Asp) Day care attendance IL5 and IL13 level Hoffjan
Toll-like receptor-2 TLR2 rs4696480 (A-16934T) Day care attendance Atopy and atopic wheezing Custovic
Toll-like receptor-9 TLR9 rs352140 Breast-feeding Food sensitization Hong
Thymic stromal lymphopoietin TSLP rs3806933 Breast-feeding Food sensitization Hong

 

CHILD STUDY TABLE G2: POLYMORPISMS IN THE TLR GENES PREVIOUSLY ASSOCIATED WITH ASTHMA AND RELATED TRAITS

Gene Gene symbol Polymorphism1 Previous association with: References
Toll-like receptor-1 TLR1 rs5743595 / rs4833095 (Asn248Ser) Atopic asthma, cytokine response / Atopic asthma Kormann / Kormann
Toll-like receptor-2 TLR2 rs4696480 / rs5743708 (Arg753Gln) Asthma /CD36, CD86 and cytokine expression in atopic dermatitis patients Eder; Niebuhr 2008, Niebuhr 2009, Mrabet-Dahbi
Toll-like receptor-4 TLR4 rs4986790 (Asp299Gly) / rs4986791 (Thr399Ile) Asthma, atopy, LPS responsiveness / Asthma, LPS responsiveness Arbour, Fageras, Yang, Werner / Werner
Toll-like receptor-6 TLR6 rs5743789 / rs5743810 (Ser249Pro) Atopic asthma, cytokine response; Atopic asthma Kormann ; Kormann
Toll-like receptor-7 TLR7 rs179008 (Gln11Leu) Asthma Moller-Larsen
Toll-like receptor-8 TLR8 rs2407992 (Leu651Leu) Asthma Moller-Larsen
Toll-like receptor-9 TLR9 rs5743836 / rs352163 Asthma / Asthma Lazarus / Smit
Toll-like receptor-10 TLR10 rs4129009 (Ile775Val) Atopic asthma, cytokine response Kormann

REFERENCES FOR METHODS TABLES

 

CHILD STUDY TABLE G1: CANDIDATE GENE STUDIES SHOWING GENE-ENVIRONMENT INTERACTIONS

  1. Bisgaard H, Simpson A, Palmer CN, Bønnelykke K, McLean I, Mukhopadhyay S, Pipper CB, Halkjaer LB, Lipworth B, Hankinson J, Woodcock A, Custovic A. Gene-environment interaction in the onset of eczema in infancy: filaggrin loss-of-function mutations enhanced by neonatal cat exposure. PLoS Med. 2008 Jun 24;5(6):e131. PMID: 18578563
  2. Bottema RW, Reijmerink NE, Kerkhof M, Koppelman GH, Stelma FF, Gerritsen J, Thijs C, Brunekreef B, van Schayck CP, Postma DS. Interleukin 13, CD14, pet and tobacco smoke influence atopy in three Dutch cohorts: the allergenic study. Eur Respir J. 2008 Sep;32(3):593-602. Epub 2008 Apr 16. PMID: 18417506
  3. Choudhry S, Avila PC, Nazario S, Ung N, Kho J, Rodriguez-Santana JR, Casal J, Tsai HJ, Torres A, Ziv E, Toscano M, Sylvia JS, Alioto M, Salazar M, Gomez I, Fagan JK, Salas J, Lilly C, Matallana H, Castro RA, Selman M, Weiss ST, Ford JG, Drazen JM, Rodriguez-Cintron W, Chapela R, Silverman EK, Burchard EG. CD14 tobacco gene-environment interaction modifies asthma severity and immunoglobulin E levels in Latinos with asthma. Am J Respir Crit Care Med. 2005 Jul 15;172(2):173-82. Epub 2005 May 5. PMID: 15879416
  4. Custovic A, Rothers J, Stern D, Simpson A, Woodcock A, Wright AL, Nicolaou NC, Hankinson J, Halonen M, Martinez FD. Effect of day care attendance on sensitization and atopic wheezing differs by Toll-like receptor 2 genotype in 2 population-based birth cohort studies. J Allergy Clin Immunol. 2011 Feb;127(2):390-397.e1-9. PMID: 21281869
  5. Hoffjan S, Nicolae D, Ostrovnaya I, Roberg K, Evans M, Mirel DB, Steiner L, Walker K, Shult P, Gangnon RE, Gern JE, Martinez FD, Lemanske RF, Ober C. Gene-environment interaction effects on the development of immune responses in the 1st year of life. Am J Hum Genet. 2005 Apr;76(4):696-704. Epub 2005 Feb 22. PMID: 15726497
  6. Hong X, Wang G, Liu X, Kumar R, Tsai HJ, Arguelles L, Hao K, Pearson C, Ortiz K, Bonzagni A, Apollon S, Fu L, Caruso D, Pongracic JA, Schleimer R, Holt PG, Bauchner H, Wang X. Gene polymorphisms, breast-feeding, and development of food sensitization in early childhood. J Allergy Clin Immunol. 2011 Aug;128(2):374-381.e2. PMID: 21689850
  7. Leynaert B, Guilloud-Bataille M, Soussan D, Benessiano J, Guénégou A, Pin I, Neukirch F. Association between farm exposure and atopy, according to the CD14 C-159T polymorphism. J Allergy Clin Immunol. 2006 Sep;118(3):658-65. Epub 2006 Jul 28. PMID: 16950285
  8. Ogbuanu IU, Karmaus WJ, Zhang H, Sabo-Attwood T, Ewart S, Roberts G, Arshad SH. Birth order modifies the effect of IL13 gene polymorphisms on serum IgE at age 10 and skin prick test at ages 4, 10 and 18: a prospective birth cohort study. Allergy Asthma Clin Immunol. 2010 Apr 20;6(1):6. PMID: 20403202
  9. Ramadas RA, Sadeghnejad A, Karmaus W, Arshad SH, Matthews S, Huebner M, Kim DY, Ewart SL. Interleukin-1R antagonist gene and pre-natal smoke exposure are associated with childhood asthma. Eur Respir J. 2007 Mar;29(3):502-8. Epub 2006 Nov 15. PMID: 17107994
  10. Reijmerink NE, Kerkhof M, Koppelman GH, Gerritsen J, de Jongste JC, Smit HA, Brunekreef B, Postma DS. Smoke exposure interacts with ADAM33 polymorphisms in the development of lung function and hyperresponsiveness. Allergy. 2009 Jun;64(6):898-904. Epub 2009 Feb 19. PMID: 19236319
  11. Rogers AJ, Brasch-Andersen C, Ionita-Laza I, Murphy A, Sharma S, Klanderman BJ, Raby BA. The interaction of glutathione S-transferase M1-null variants with tobacco smoke exposure and the development of childhood asthma. Clin Exp Allergy. 2009 Nov;39(11):1721-9. PMID: 19860819 Simpson A, John SL, Jury F, Niven R, Woodcock A, Ollier WE, Custovic A. Endotoxin exposure, CD14, and allergic disease: an interaction between genes and the environment. Am J Respir Crit Care Med. 2006 Aug 15;174(4):386-92. Epub 2006 Apr 13. PMID: 16614348
  12. Smit LA, Siroux V, Bouzigon E, Oryszczyn MP, Lathrop M, Demenais F, Kauffmann F; Epidemiological Study on the Genetics and Environment of Asthma, Bronchial Hyperresponsiveness, and Atopy (EGEA) Cooperative Group. CD14 and toll-like receptor gene polymorphisms, country living, and asthma in adults. Am J Respir Crit Care Med. 2009 Mar 1;179(5):363-8. Epub 2008 Dec 18. PMID: 19096003
  13. Tantisira KG, Silverman ES, Mariani TJ, Xu J, Richter BG, Klanderman BJ, Litonjua AA, Lazarus R, Rosenwasser LJ, Fuhlbrigge AL, Weiss ST. FCER2: a pharmacogenetic basis for severe exacerbations in children with asthma. J Allergy Clin Immunol. 2007 Dec;120(6):1285-91. Epub 2007 Nov 5. PMID: 17980418
  14. Virta M, Pessi T, Helminen M, Seiskari T, Kondrashova A, Knip M, Hyöty H, Hurme M. Interaction between CD14-159C>T polymorphism and Helicobacter pylori is associated with serum total immunoglobulin E. Clin Exp Allergy. 2008 Dec;38(12):1929-34. PMID: 19037967 Wang Z, Chen C, Niu T, Wu D, Yang J, Wang B, Fang Z, Yandava CN, Drazen JM, Weiss ST, Xu X. Association of asthma with beta(2)-adrenergic receptor gene polymorphism and cigarette smoking. Am J Respir Crit Care Med. 2001 May;163(6):1404-9. PMID: 11371409
  15. Wu AC, Lasky-Su J, Rogers CA, Klanderman BJ, Litonjua AA. Fungal exposure modulates the effect of polymorphisms of chitinases on emergency department visits and hospitalizations. Am J Respir Crit Care Med. 2010 Oct 1;182(7):884-9. Epub 2010 Jun 10. PMID: 20538957 Zambelli-Weiner A, Ehrlich E, Stockton ML, Grant AV, Zhang S, Levett PN, Beaty TH, Barnes KC. Evaluation of the CD14/-260 polymorphism and house dust endotoxin exposure in the Barbados Asthma Genetics Study. J Allergy Clin Immunol. 2005 Jun;115(6):1203-9. PMID: 15940135
  16. Zhang G, Khoo SK, Laatikainen T, Pekkarinen P, Vartiainen E, von Hertzen L, Hayden CM, Goldblatt J, Mäkelä M, Haahtela T, Le Souëf PN. Opposite gene by environment interactions in Karelia for CD14 and CC16 single nucleotide polymorphisms and allergy. Allergy. 2009 Sep;64(9):1333-41. Epub 2009 Feb 14. PMID: 19222419.

CHILD STUDY TABLE G2: CANDIDATE GENE STUDIES SHOWING GENE-ENVIRONMENT INTERACTIONS

  1. Arbour NC, Lorenz E, Schutte BC et al. TLR4 mutations are associated with endotoxin hyporesponsiveness in humans. Nat Genet. 2000; 25:187-191. Eder W, Klimecki W, Yu L et al. Toll-like receptor 2 as a major gene for asthma in children ofEuropean farmers. J Allergy Clin Immunol. 2004; 113:482-488.
  2. Fageras BM, Hmani-Aifa M, Lindstrom A et al. A TLR4 polymorphism is associated with asthma and reduced lipopolysaccharide-induced interleukin-12(p70) responses in Swedish children. J Allergy Clin Immunol. 2004; 114:561-567.
  3. Kormann MS, Depner M, Hartl D et al. Toll-like receptor heterodimer variants protect from childhood asthma. J Allergy Clin Immunol. 2008; 122:86-92.
  4. Lazarus R, Klimecki WT, Raby BA et al. Single-nucleotide polymorphisms in the Toll-like receptor 9 gene (TLR9): frequencies, pairwise linkage disequilibrium, and haplotypes in three U.S. ethnic groups and exploratory case-control disease association studies. Genomics. 2003; 81:85-91.
  5. Moller-Larsen S, Nyegaard M, Haagerup A et al. Association analysis identifies TLR7 and TLR8 as novel risk genes in asthma and related disorders. Thorax. 2008; 63:1064-1069.
  6. Mrabet-Dahbi S, Dalpke AH, Niebuhr M et al. The Toll-like receptor 2 R753Q mutation modifies cytokine production and Toll-like receptor expression in atopic dermatitis. J Allergy Clin Immunol. 2008;121:1013-1019.
  7. Niebuhr M, Langnickel J, Draing C et al. Dysregulation of toll-like receptor-2 (TLR-2)-induced effects in monocytes from patients with atopic dermatitis: impact of the TLR-2 R753Q polymorphism. Allergy. 2008; 63:728-734.
  8. Niebuhr M, Langnickel J, Sigel S, Werfel T. Dysregulation of CD36 upon TLR-2 stimulation inmonocytes from patients with atopic dermatitis and the TLR2 R753Q polymorphism. Exp Dermatol. 2009.
  9. Smit LA, Siroux V, Bouzigon E et al. CD14 and toll-like receptor gene polymorphisms, country living, and asthma in adults. Am J Respir Crit Care Med. 2009; 179:363-368. Werner M, Topp R, Wimmer K et al. TLR4 gene variants modify endotoxin effects on asthma. J Allergy Clin Immunol. 2003;112:323-330.
  10. Yang IA, Barton SJ, Rorke S et al. Toll-like receptor 4 polymorphism and severity of atopy in asthmatics. Genes Immun. 2004; 5:41-45.

Epigenetic mechanisms mediate the effects of early life environments to persistently alter the expression of genes and functions of cells critical to the development of allergic diseases and asthma.

 

Despite the fact that many environmental exposures have been linked with childhood asthma, we are just beginning to unravel how environmental exposures modify the function of asthma susceptibility genes. Epigenetics has emerged as a critical bridge to understanding the causes of complex diseases such as allergies and asthma. Epigenetic research focuses upon chromatin, which consists primarily of DNA and histone proteins, and functions as the principal template for the creation and maintenance of 'epigenetic marks'.

 

Core epigenetic features such as DNA methylation and histone modifications are likely to play critical roles, given the relationship between epigenetic markers and gene function, the evidence that environmental factors can elicit epigenetic changes, and that epigenetic mechanisms regulate the expression of asthma susceptibility genes and cells involved in the development of allergy. Modulation of epigenetic processes has been shown to influence models of allergic disease in mice, and exposure to polycyclic aromatic hydrocarbons in utero affects DNA methylation and risk of asthma in childhood. Epigenetics is also emerging as an integral component of the mind-body axis, with the early-life social environment influencing the expression and DNA methylation of genes into adulthood.

 

Specific Testable Hypothesis Related to Epigenetics:

  1. Oxidizing exposures influence the inflammatory epigenome expressed at 3 months. This analysis will involve prenatal exposures, maternal diet and exposure index at 3 months including medications.
  2. Epigenetic mechanisms regulate the expression of genes and functions of cells critical to the development of allergic diseases and asthma (see Gene-Environment REEGLE project).
  3. Significant changes in DNA methylation observed between cord blood and 1 year blood are correlated with environmental exposures including traffic-related air pollution, stress, environmental tobacco smoke and phthalate exposure and there may be a link to the flora colonizing infant intestinal tract. This is the specific direction intended for REEGLE.  An analysis plan, once the samples are selected and the methylation arrays are done, needs to be developed.
  4. FOXP3 DNA methylation is altered in Treg cells isolated from umbilical cord blood of babies born to allergic mothers. Two UK researchers have already shown this in a small Kingston birth cohort and replicating in CHILD would be valuable 

  5. Changes in DNA methylation observed between cord blood and 1 year blood persist at later time-points, and the most persistent methylation patterns are associated with chronic environmental exposures. Timeline – required 5 year blood – 2016-17.
  6. Environmental exposures in utero or early life lead to permanent changes in DNA methylation patterns which will manifest in later years as an allergic outcome. Timeline: This requires 3 year and 5 year outcomes.

 

 

The novelty of the CHILD study:

In the CHILD Study researchers will investigate epigenetic dysregulation, interrogating DNA methylation across biologically relevant CpG loci as a reliable and quantifiable effect of environmental factors on the genome. This will enable CHILD to advance knowledge of the role that epigenetics play in mediating effects of environmental exposure, including stress, on the development and progression of asthma and allergic diseases, and may allow discovery of early DNA methylation-based biomarkers for asthma development.

 

Few publications have looked at age-dependent changes in DNA methylation in the same individual and this presents an unprecedented opportunity for epigenetics as it allows the interrogation of epigenetic marks across early-life development and their association with environmental exposures and health outcomes. The pediatric age range might be particularly sensitive to changes in DNA methylation, highlighting the need for longitudinal measurements. Perhaps many of the dramatic changes in the epigenome found between fetal and adult tissues are set in early life, consistent with the overarching hypothesis of CHILD.

 

In addition, CHILD researchers and others have recently found that the level of methylation at a substantial fraction of CpG loci is correlated to nearby SNPs and differ between ethnic populations, thus raising the tantalizing possibility of interplay of genetics and epigenetics in the etiology of human disease.

EPIGENETICS - Power / Sample Size

As examples of potential studies of DNA methylation and gene expression, we will test the relationships between mRNA expression of the glucocorticoid receptor (GR) and the β- adrenergic receptor and the mean levels of methylation of the promoter regions of the GR and β- adrenergic receptor genes as well as of 10-15 specific CpG sites within the proposed transcription factor binding regions of these genes. Since we will be testing for multiple CpG sites per gene, we set an alpha of 0.005 for significance and performed simulations with 600 and 750 subjects.

 

Correlations between the DNA methylation status and the gene expression were specified in a range of -0.9 to -0.1, since a negative correlation between the variables is expected. CHILD researcher calculations show that there will be greater than 80% power to detect correlations greater than -0.15. For example, a sample size of 600 achieves over 98% power to detect a difference between the null hypothesis (correlation of 0.0) and an r of -0.2 using a two-sided hypothesis test with a significance level of 0.005.

EPIGENETICS Methods

For the genome-wide assays, bisulfite converted DNA will be processed for hybridization to the Illumina 450K Infinium Methylation BeadChip array. Currently, the laboratory has performed some 1000 assays using this technology, establishing its feasibility for the CHILD project. This platform interrogates >450,000 CpG methylation sites per sample and covers all designable RefSeq genes, with CpG Island shores, non-island CpGs, CpG islands outside of coding regions and miRNA promoter regions represented. In addition, non-CpG sites found to be differentially methylated in human stem cells are included, as well as regions identified in GWAS studies to be disease-associated and sites identified as differentially methylated in tumor versus normal studies. The assay comfortably allows a throughput of 96 samples per week.

 

Using statistical approaches based on permutation testing and controlling for blood composition, candidate CpG loci associated with variables available from the CHILD cohort will be identified and confirmed by pyrosequencing.

 

For the candidate gene approach, selected regions of approximately 200 basepair length in the candidate gene regions emerging from the array-based studies and in literature-derived candidate genes will be interrogated for quantitative measurements of CpG methylation at single nucleotide resolution using pyrosequencing. The selection of regions will focus on functionally relevant regions such as binding sites for other transcription factors and regions adjacent to genetic variants.

 

 

 

The CHILD Study, with its wealth of longitudinal environmental, immunological, physiologic and genetic data, will allow both traditional and novel strategies of data analysis to test multiple hypotheses. CHILD will also allow testing of new hypotheses which emerge as science advances, utilizing the high quality questionnaire, environmental and biological sample databases in CHILD.

 

The strengths of the CHILD Study include these multiple avenues of investigation, integrating the diversity of proposed investigations developed by task-specific Scientific Working Groups, allowing the testing of a wide range of specific hypotheses related to the origins of allergy and asthma.

 

The sub-hypotheses remains a 'work-in-progress' as the extensive list of testable hypotheses and related analyses proposed by the Scientific Working Groups within the CHILD study will be refined, prioritized, and reviewed regularly to track progress and monitor deliverables arising from the study. Note that:

  • interactions between different sectors of CHILD will be further developed in proposals for analyses (for example, psychosocial environment and markers of innate and adaptive immunity, already explored in the mini-CHILD cohort)
  • some of the proposed analyses will require the maximum available sample size which is now limited to 3000-3500 children depending on age, and so cannot be undertaken until all children have reached the appropriate milestones
  • some analyses will be most cost-effective conducted as case-control analyses, meaning selection of those samples to analyse will of necessity be deferred until case status (intermediate or primary outcome) is known
  • some of the specific hypotheses listed are so novel that no data exist for determining power
  • given the finite number of children in the cohort, some analyses may not reach traditional statistical significance, but nevertheless will provide first insights regarding a number of biological outcomes which will guide future studies

 

 

CHILD Logo and Tag