Prematurity and early life programming
View/ Open
Piyasena2016.docx (1.895Mb)
Date
02/07/2016Author
Piyasena, Chinthika
Metadata
Abstract
Preterm infants are at increased risk of cardiometabolic and neurodevelopmental
disorders in later life. The typical postnatal growth pattern of failure to achieve the
equivalent of a normal fetal growth rate, followed up by catch-up growth, altered
adiposity and altered hypothalamic-pituitary-adrenal axis (HPA) activity may be
predisposing factors. Potential mechanisms that may mediate such programmed
effects include altered DNA methylation and faster telomere attrition.
A prospective cohort of 46 very preterm (25+2 to 31+5 weeks’ gestation, mean 28.6)
and 40 full term (38+3 to 42+2 weeks’ gestation, mean 40.2) infants was established
to investigate potential mechanisms. Infants were studied at birth, term equivalent
age, 3 months and 1 year corrected for prematurity. At all time points, linear growth
and body composition (by densitometry) were measured and buccal (epithelial) cells
was collected for measurement of DNA methylation (5mC) and relative telomere
length.
Compared with full term infants, preterm infants were lighter (p < 0.001) and had a
smaller head circumference (p < 0.05) at all time-points and were shorter at term
equivalent (p < 0.001) and 3 months corrected age (p = 0.002). Preterm infants also
had greater percentage body fat at term equivalent age (mean difference = 5.5%, p <
0.001), which normalised by 3 months corrected (mean difference = 0.9%, p = 0.4).
Preterm infants had a blunted salivary cortisol response (mean difference 0.4 μg/dL,
p = 0.02) to a stressor (physical examination) at 3 months compared to term infants
at this age, suggesting altered activity of the HPA axis.
5mC is fundamental in the control of expression of imprinted genes involved in fetal
growth. Notably, a number of studies in humans exposed to an adverse environment
in early life have demonstrated altered 5mC at the differentially methylated regions
(DMRs) controlling the expression of the key fetal growth factor insulin like growth
factor 2 (IGF2) and at the linked H19 imprinting control region (H19 ICR). At birth,
preterm infants had a significant decrease in 5mC at DMR2 compared with term
infants at birth (β = –11.5, p < 0.001) and compared with preterm infants at term
equivalent age (mean difference = -7.4, p = 0.01). By term equivalent age, preterm
infants had decreased 5mC at both DMR2 (β = –2.8, p = 0.01) and the H19 ICR (β =
–2.3, p = 0.048) compared with term infants at birth, although this difference
disappeared at 1 year corrected. Although research has suggested that catch up
growth may confer an unfavourable metabolic phenotype, poor initial weight gain
can associate with worse cognitive outcome. A pathway was established for
obtaining advanced magnetic resonance images of the preterm brain. 5mC at H19
ICR and DMR2 in buccal DNA showed no association with measures of white matter
microstructure or whole brain volumes.
Term infants demonstrated telomere lengthening over the first year of life (mean
difference = -0.3, p = 0.02). There was no significant change in telomere length over
the first year of life in preterm infants (mean difference = 0.2, p = 0.34). However, as
preterm infants at term equivalent age had longer telomeres compared to term infants
at birth (β = 0.6, p < 0.001), ultimately there were no differences between the term
and the preterm groups at 1 year corrected age (β = 0.3, p = 0.07).
The DNA modification 5-hydroxymethylcytosine (5hmC) is a stable modification in
its own right and is also thought to be an intermediate step in DNA demethylation.
5hmC is abundant in the placenta but has not been studied in the context of fetal
programming. Additionally, previous research using methods such as bisulphite
conversion would not have discriminated between 5mC and 5hmC and therefore the
role of 5mC may not have been accurately measured. To study the relationship
between 5mC, 5hmC and fetal growth, gene expression of candidate imprinted and
non-imprinted genes in full term placental samples from the Edinburgh Reproductive
Tissue BioBank was analysed. 5mC and 5hmC within the IGF2/H19 and KvDMR
(controlling CDKN1C) loci was estimated using chemical capture and
immunoprecipitation techniques that discriminate between modifications.
Relationships between the expression of IGF2 (r = 0.3, p = 0.02) and CDKN1C (r =
-0.3, p = 0.01) and birth weight across the normal range were found and in keeping
with the known action of these genes. 5mC at IGF2 DMR0 (β = 0.3, p = 0.02) and
KvDMR (β = 0.3, p = 0.02) and 5hmC at H19 gene body (β = 0.2, p = 0.04)
associated with birth weight.
Thus, DNA modifications at imprinted DMRs may modulate environmental
influences on fetal growth across the normal range. DNA methylation at IGF2/H19
can be influenced by early life events. It remains to be seen whether any changes are
present later in childhood and whether they associate with risk factors for the
metabolic syndrome.