br Materials and Methods br Results br Discussion

Materials and Methods
Results
Discussion In this study, we identified several key developmental and stem cell regulatory genes as potential biomarkers of arsenic exposure in the fetal placenta. Our results also revealed extensive sexual dimorphism in the associations between placental gene expression and both in utero arsenic exposure and infant birth weight. Particularly, we found that expression of the HH pathway component, GLI3, in female placentae was associated with both arsenic exposure and infant birth weight. Combined with our previous results implicating GLI3 repressor protein as a key arsenic target in cultured ahr pathway (Fei et al., 2010), these results suggest that GLI3 may be a pivotal signaling node affected by arsenic (Fig. 4B). With respect to determining the molecular mechanisms underlying arsenic\'s effects on children\'s health, our study has some limitations that should be acknowledged. Firstly, this is a cross-sectional study, using single measurements of U–As at mid-gestation, and placental gene expression and infant weight at delivery. Therefore, this study was not designed to identify cause-and-effect relationships. Additionally, the chronological gap between the urinary arsenic measurement and analyses of birth weight and placental gene expression may have led to inaccurate assessment of arsenic exposure for some individuals, although previous studies show that an individual\'s urinary arsenic concentration is generally a reliable indicator of long-term exposure levels (Kile et al., 2009; Navas-Acien et al., 2009). Finally, our study was performed on a relatively small cohort and so was statistically underpowered to measure the direct relationship between arsenic exposure and birth weight, although the trend was similar to previous reports. Despite these limitations, this study highlights possible genes and pathways mediating arsenic\'s effects on fetal development and growth, providing an important basis for future studies to confirm the causative relationships. Arsenic exposure levels in our population were relatively low and the gene expression changes we observed were also relatively subtle. Nonetheless, even small variations in the expression of key developmental pathway genes in utero could still impact fetal development and growth by influencing placental function. Moreover, since arsenic has been shown to readily pass across the placenta (Concha et al., 1998), it is likely that the gene expression changes we observed in the fetal placenta also occur in other fetal tissues, and directly influence their development and growth. In our cohort, we observed arsenic-associated reductions in the expression of the HH pathway target GLI1, and the stem cell regulator, POU5F1. POU5F1 (also called OCT4) is a critical player in maintaining pluripotency of embryonic stem cells, and its down-regulation is necessary for differentiation of the trophectoderm, which ultimately forms the placenta (Nichols et al., 1998). Interestingly, residual expression of POU5F1 is detected throughout the first trimester and in term placenta tissues, and reduction in this residual expression is associated with gestational trophoblastic diseases (Zhang et al., 2008). In the fetus, early loss of POU5F1 expression leads to preimplantation lethality (Nichols et al., 1998), and during later development, POU5F1 is essential for convergent extension and primordial germ cell survival (Kehler et al., 2004; DeVeale et al., 2013). The loss of one copy of this gene, causing a reduction in expression of 30–40%, can induce differentiation of embryonic stem cells in vitro (Nichols et al., 1998). Thus, an 8% reduction in POU5F1 expression, which we observe in response to arsenic exposure between the 25th and 75th quartiles of U–As levels, may have multiple effects on fetal health. Relatively little is known about the role of GLI1 in placenta biology. However, in mice, several components of the HH pathway, including all three HH ligands; SMO; PTCH1; and the transcription factors GLI2 and GLI3, are expressed abundantly in fetal placenta (Pan et al., 2015), and disruption of HH activity causes severe placental defects, suggesting essential roles in normal placental development and possibly pregnancy maintenance (Pan et al., 2015; Jiang and Herman, 2006). Gli1 homozygous null mice have reduced body weight, increased postnatal lethality, and defects in T-cell development, while heterozygotes are of normal weight but have reduced bone mass, reflecting the importance of normal levels of GLI1 for bone homeostasis and immune function (Kitaura et al., 2014; Drakopoulou et al., 2010). Moreover, a common single nucleotide polymorphism (SNP) in human GLI1, which reduces protein activity by around 50%, is associated, in both homozygous and heterozygous states, with increased risk of inflammatory bowel disease (Lees et al., 2008). Therefore, subtle changes in GLI1 expression, such as the 18% reduction we observe in response to arsenic exposure between the 25th and 75th quartiles of U–As levels, may have important consequences for the fetus.