Dynamics of DNA methylation at IGF2 in preterm and term infants during the first year of life: an observational study.

BACKGROUND: Preterm infants are at increased risk of cardiometabolic disease in later life. Extrauterine growth restriction, catch-up growth, altered adiposity, and abnormal hypothalamic-pituitary-adrenal axis activity could be predisposing factors. Altered DNA methylation (5-methylcytosine, 5mC) might be one underlying mechanism. We hypothesised that preterm infants have altered 5mC at the linked differentially methylated region 2 (DMR2) of IGF2 and the H19 imprinting control region (H19 ICR) compared with term infants over the first year of life.

METHODS: We recruited 46 preterm (range 25 weeks + 2 days' gestation to 31 + 5, mean 28 + 6) and 40 term infants (38 + 3 to 42 + 2 weeks' gestation, mean 40 + 2). Anthropometric variables including body composition were measured at term age and 3 months corrected age with air displacement plethysmography and at 1-year-corrected age with skin-fold thickness. Salivary cortisol was measured at 3 months corrected age after the physical examination. Percentage methylation (%5mC) was analysed with pyrosequencing on buccal DNA. Statistical analysis used Student's t test and multivariate linear regression.

FINDINGS: Preterm infants demonstrated growth deficit early in postnatal life but had greater percentage body fat at term age (β=5·73, p<0·001), but not at 3 months (β=-0·28, p=0·82). Compared with term infants, preterm infants had a blunted cortisol response to physical examination (mean difference 0·38 μg/dL, p=0·024). At birth, preterm infants had a significant decrease in %5mC at DMR2 compared with term infants at birth (β=-11·48, p<0·001) and compared with preterm infants at term-corrected age (t=3·13, p=0·01). By term-corrected age, preterm infants had decreased %5mC at both DMR2 (β=-2·84, p=0·013) and the H19 ICR (β=-2·31, p=0·048) compared with term infants at birth, although this difference disappeared at 1 year. Social deprivation was independently associated with decreased %5mC at DMR2 at birth (β=-1·73, p=0·006) and term-corrected age (β=-0·86, p=0·016) but not at 1 year (β=-0·89, p=0·07).

INTERPRETATION: Our results show that decreased %5mC accompanies the early growth deficit in preterm infants. The marked reduction in %5mC at IGF2 DMR2 in preterm infants at birth compared with term-age supports existing evidence that imprinting at secondary regions is established after fertilisation, whereas imprinting is established during gametogenesis at primary regions (H19 ICR). Both regions might be susceptible to early life stressors such as preterm birth and social deprivation.

FUNDING: Chief Scientist Office of the Scottish Government.