The demographics of the 186 infants in our population that were studied are provided in . Based on the intended composition of the cohort, the population is over-represented by small and large for gestational age infants (SGA and LGA, respectively). All infants enrolled were considered at or near term, with a mean gestational age of 39.1 weeks. The racial profile of our population is mixed, with 10% of mothers reporting an African-American race, and over 20% a race other than caucasian or African American. Only 4% of participants reported smoking tobacco during pregnancy, and less than 1% reported alcohol or illegal drug use during pregnancy. The sociodemographic characteristics of the population is also mixed, with 54% of the population utilizing public health insurance. Also included in are the descriptive statistics of the summary scores for the 10 NNNS measures examined in this study.
Population Demographics, Clinical Characteristics, and NNNS Summary Scores.
Quantitative bisulfite pyrosequencing was used to determine the DNA methylation status of a CpG island region in the promoter of the HSD11B2
gene previously shown to exhibit differential methylation in human placenta tissue in our sample of 186 placentas 
. To determine the functional significance of the variation in methylation in these samples, we also performed gene expression analysis using qRT-PCR in a subset of 95 placentas examined. A moderate, statistically significant negative correlation (ρ
−0.24, P<0.02) was observed between mean extent of HSD11B2
methylation across the 4 CpG loci examined and mRNA expression of HSD11B2
(). We have previously demonstrated a relationship between infant birth weight status and DNA methylation of the glucocorticoid receptor promoter 
and so examined the association between various measures of infant growth and maternal and infant clinical characteristics and overall mean methylation of the HSD11B2
gene promoter region. presents the bivariate correlations between mean extent of HSD11B2
methylation and each of these characteristics. Moderate, statistically significant negative, correlations were observed between infant birthweight, and ponderal index (ratio of weight for length) and HSD11B2
methylation, demonstrating smaller and leaner/thinner infants had increased levels of HSD11B2
methylation. This was clearly shown by comparison of the extent of HSD11B2
methylation in intrauterine growth restricted vs. non-IUGR infants, where IUGR infants demonstrated a significantly greater extent of HSD11B2
0.007, ). More modest, yet statistically significant negative correlations were observed between HSD11B2 methylation extent and infant length and gestational age.
Scatterplot of the correlation between mean HSD11B2 methylation extent and normalized HSD11B2 mRNA expression.
Plots of the correlations between HSD11B2 mean methylation extent and infant characteristics.
The relationship between these measures of growth and HSD11B2 methylation extent, controlled for infant gender, maternal age, and gestational age are provided in . There is a consistent negative association between birthweight, ponderal index and gestational age and the extent of methylation in the HSD11B2 promoter, but the association between infant length and methylation is attenuated in these adjusted models. The relationship is particularly strong, consistent with the bivariate tests, between IUGR status and methylation. Taken together, these results demonstrate that specific measures of intrauterine growth are associated with DNA methylation of this gene in the placenta.
Generalized Linear Models of the Association between HSD11B2 Methylation Extent and Maternal and Infant Characteristics Controlled for Confounders.
The relationship between mean methylation extent of the HSD11B2
gene promoter and each of the individual NNNS summary scores was examined using a Spearman rank correlation, the results of which are tabulated in . Infant quality of movement scores showed a significant negative correlation, suggesting that increased HSD11B2
methylation was associated with poorer quality of movement (ρ
0.01). Infant hypertonicity also showed a negative trend in its association with HSD11B2 methylation (ρ
0.06), while infant attention showed a positive trend in its association with HSD11B2
methylation, with increasing methylation related to increasing infant attention scores (ρ
0.08). depicts these relationships. The other scores were not associated with HSD11B2
methylation at the P<0.1 level.
Bivariate Correlations Between HSD11B2 Mean Methylation Extent and NICU Network Neurobehavioral Scales Summary Scores.
Plots of the relationships of HSD11B2 mean methylation and selected NICU Network Neurobehavioral Scales.
To examine if the association between NNNS outcomes and HSD11B2 methylation was independent of infant growth, gender, and maternal age, we utilized multivariable generalized linear regression to model NNNS outcomes (quality of movement, hypertonicity, and attention) with HSD11B2 methylation alone, or controlling for these additional factors. The results of these models including the effect sizes (ß values) and statistical significance are shown in . As expected from the univariate correlation analyses, in the unadjusted model, an increase in methylation of 10 fold (i.e. 1 log) of the HSD11B2 promoter region was significantly associated with a decrease in infant quality of movement of 1.3 units. Controlling for infant growth, gender, and maternal age attenuated slightly the effect, of HSD11B2 methylation on quality of movement, with a 10-fold increase of methylation now associated with a significant decrease in quality of movement of 1.2 units, independent of growth, gender and maternal age. Interestingly, each year of maternal age was associated with a small but significant increase in quality of movement independent of the other factors. On the other hand, a 10-fold increase in methylation showed a trend for an increase in infant attention score of 1.89 units, which was attenuated and not significant when controlled for infant growth, as well as infant gender and maternal age. Hypertonicity decreased by nearly 1 unit with a 10-fold increase in HSD11B2 methylation, a result that also was attenuated and not significant when confounders were included in the model.
Generalized Linear Models of the Association between HSD11B2 Methylation Extent and NICU Network Neurobehavioral Scales Quality of Movement, Attention, and Hypertonicity Scales.