Transcript Fellows seminar 9-19-2014

Low birth weight
Zahra N. Sohani
Supervisor: Dr. Sonia Anand
Outline of presentation
• Background & brief methods of project
• Issues to consider for analysis
Low birth weight & type 2 diabetes –
Epidemiological evidence
 Hales & Barker reported an increase in incidence
of IGT and T2DM in children born small
 Systematic review of 31 studies (n=6090 diabetes
cases) showed a 25% reduction in odds of T2DM
per kg of weight gained
 Another study comparing birth weight <2.5 kg
with ≥2.5 kg found a 32% increase in odds of
T2DM for those born with low birth weight
The hypotheses
Impaired beta cell function
 fetal period is critical for
pancreatic development
 impairment in beta cell
development leads to reduced
insulin secretion, which results in
reduced fetal growth
 Beta cell genes are also
associated with birth weight
 Studies of human fetuses with
growth restrictions have reported
reduced pancreatic endocrine
cell mass
Insulin resistance
 in utero perturbations cause the
fetus to develop peripheral insulin
resistance, to divert nutrients to
essential organs
 results in permanent reductions in
skeletal muscle glucose
transporter number and/or
function, causing extensive insulin
secretion from the pancreas early
in childhood, but eventual beta
cell exhaustion
 Children with IUGR show greater
insulin response than those born
with normal birth weight
Is there support for
epigenetics?
 Methylation is one of the commonly studied
epigenetic marks
 addition of a methyl group to the 5-carbon position of a
cytosine residue located next to guanine
1. methylation can lead to the binding of methylated
CpG binding proteins and transcriptional repressors to
the methyl group which block transcription factor
access
2. presence of the methyl group itself can inhibit binding
of necessary transcription factors to initiate the process
via steric hindrance
 DNA methylation is erased throughout the genome
and then re-established during embryogenesis &
gametogenesis
 Allows for ‘soft inheritance’
 Human studies
 Inheritance of MLH1 epimutation (linked to certain
cancers)
 Germline epimutation of MSH2 in a family with
hereditary nonpolyposis colorectal cancer
 Animal models
 model of rats demonstrated persistence of LBW for
three generations after re-introduction of normal
nourishment to offsprings of malnourished mothers
Questions
 Are there differentially methylated regions
associated with birth weight?
 Is there a difference in these regions between
South Asians and white Europeans?
Cardio-metabolic profile in
South Asians
 South Asians have greater visceral adiposity, and
experience metabolic abnormalities at lower a
BMI and younger age than Europeans
 South Asian babies show a thin-fat phenotype
 South Asian newborns are born lighter even after
generations in ‘western’ countries
Brief methods
 Phenotypes: birth weight, birth length, and percent
body fat
 Genotypes: SNPs involved in beta cell function,
insulin resistance, birth weight, adiposity, birth
length, percent body fat, type 2 diabetes, and
obesity in newborns and adults
 Methylation/Expression: Differentially methylated
regions for SNPs involved in beta cell function, insulin
resistance, birth weight, adiposity, birth length,
percent body fat, type 2 diabetes, and obesity
 Look for gene expression differences if any DMRs are
significantly associated with phenotype
Some current evidence
 Three significant methylation studies investigating
birth weight
 Engel (2014) report Bonferroni-corrected associations for
19 CpGs with birth weight
 Tarun (2012) identified 23 genes for which methylation
collectively explained 70-87% of the variance in birth
weight
 Gordon (2012) found methylation in 8 genes to be
associated with birth weight after correction for multiple
testing
Issues to consider
 Beta values versus M-values
Du BMC Bioinformatics 2010
 Titration experiment of two samples known to be
differentially methylated: 100:0, 90:10, 75:25,
50:50, 0:100
 Methylation profile as established by beta values
and M-values should correlate to titration profile
 Beta-value method has a direct biological
interpretation - it corresponds roughly to the
percentage of a site that is methylated
 M-value method is more statistically valid, but
does not have an intuitive biological meaning