Sex Steroid Hormone Levels and Reproductive Development of Eight-Year-Old Children following In Utero and Environmental Exposure to Phthalates

In utero exposure to phthalates may adversely affect reproductive development in children due to the anti-androgenic properties of the pthalates. Accordingly, we aimed to determine the effects of in utero and environmental phthalate exposure on the reproductive development of eight-year-old children. We recruited 180 children in central Taiwan during November 2001 and followed them until August 2009 when all children became eight years old. Birth outcomes were collected. Bone age, hormone concentrations, and reproductive developmental stages were determined. Phthalate metabolite levels, including mono-2-ethylhexyl phthalate [MEHP], mono-n-butyl phthalate [MnBP], and mono-benzyl phthalate [MBzP], were assessed. No significant gender differences were found in in utero phthalate exposure. Maternal urinary levels of phthalate metabolites did not correlate significantly with birth outcomes, physical characteristics, and reproductive hormones of the eight-year-old children. Regarding the urinary phthalate metabolite levels of the eight-year-old children, MEHP correlated significantly with serum progesterone levels. MEHP levels in girls correlated significantly with serum progesterone levels. MnBP correlated significantly with serum FSH in all children. In girls, MnBP correlated with serum FSH, and MBzP correlated with serum progesterone and FSH levels. Urinary phthalate metabolite levels did not correlate with female developmental stages or the development of female reproductive organs. Phthalate metabolites did not correlate with the physical characteristics and reproductive hormones in boys. Therefore, environmental exposure to phthalates, as determined by urinary phthalate metabolite levels of eight-year-old children, may affect reproductive hormone levels in children, indicating that further studies on the environmental health effects of phthalates are warranted.


Introduction
Phthalates are synthetic chemicals used primarily as plasticizers and chemical additives for various applications. Phthalates are released continuously from such products and are ubiquitous in the environment owing to their chemical properties (i.e., no chemical bonds formed between phthalates and plastics or additives). Consequently, human exposure to phthalates is widespread and common, particularly since phthalates can be absorbed through the skin, inhaled, ingested, or directly administered [1,2].
Recently, biomonitoring studies have demonstrated relatively high urinary levels of phthalate metabolites in the general population of Western countries [2][3][4][5]. Furthermore, concentrations in children, who are more sensitive to exogenous insults, were found to be even higher [1,2,6].
This exposure is of great concern, as phthalates have been suspected to have anti-androgenic and estrogenic effects, which may alter the reproductive development of children [7,8]. Indeed, toxicological evidence suggests that some phthalates, including butyl benzyl phthalate (BBzP), di-n-butyl phthalate (DnBP), and di-(2-ethylhexyl) phthalate (DEHP), may alter or mimic estradiol (E2) in vivo and in vitro [8]. Animal studies indicate that gestational phthalate exposure is associated with adverse health outcomes, including disruption of endocrine and reproductive function and development [1]. Furthermore, it was suggested that most reproductive effects of phthalates are not exerted by phthalates per se, but rather their metabolites [9]. Recently, a study by Huang et al. [10] demonstrated that subjects with a glutathione S-transferase M1 null mutant genotype had not only higher urinary levels of total urinary mono-ethylhexyl phthalate (MEHP), a metabolite of DEHP, but also a significantly higher risk for estrogen-dependent diseases [10]. Additionally, studies on newborns have demonstrated associations between prenatal and early postnatal phthalate exposure with shorter anogenital distance, lower serum testosterone levels, and alterations in birth outcomes [11][12][13]. However, evidence regarding the effects of phthalates on human reproduction, particularly the effects of in utero exposure on the reproductive development of pre-pubertal children, is still very limited [14].
In the present study, we aimed to examine the association between in utero and environmental exposure to phthalate metabolites and the reproductive development of pre-pubertal (i.e., age 8 years) children in Taiwan. The correlations between in utero exposure levels and birth outcomes, reproductive hormone levels, and female sexual characteristics were analyzed.

Participants
The present study was a prospective follow-up study of 130 children and their mothers, who were recruited, and were followed from November 2001 to August 2009. This study of the general population in central Taiwan served as pilot for a nation-wide Taiwan Maternal and Infant Cohort Study (TMICS). All children were eight years old at the time of follow-up. Subject recruitment was previously described. This study was approved by the ethics review committee of the National Health Research Institutes in Taiwan, and parental written informed consent was obtained for each participant.

Data and specimen collection
Participants' characteristics at birth and age eight years were determined from records. Data on birth outcomes included sex, gestational age, birth length, birth weight, head circumference, and breast circumference. At age eight years, height, weight, body mass index (BMI), systolic blood pressure (SBP), and diastolic blood pressure (DBP) were recorded.
At follow-up, bone age (BA) was determined by examining left hand radiographs using the method of Greulich and Pyle [15], and the BA/chronological age (CA) ratio was calculated. Additionally, eight-hour fasting blood samples were obtained from each participant, immediately centrifuged, and the serum was separated and frozen at 270uC for further analysis.

Reproductive hormone measurements
Serum total testosterone (TT) and progesterone levels were measured using a solid-phase, competitive chemiluminescent enzyme immunoassay (Immulite 2000 Systems Analyzers, Siemens Medical Solutions Diagnostics, Deerfield, Illinois). The sensitivity for TT was 1.5 ng/mL, and for progesterone was 0.1 ng/mL (0.3 nmol/L). Free testosterone levels were determined by a Coat-A-Count Free Testosterone 125 I radioimmunoassay kit purchased from Diagnostic Products Corp. (Los Angeles, California), which had a sensitivity of 0.15 pg/mL. Serum estradiol (E2) levels were measured using a radioimmunoassay kit (Diagnostic Systems Laboratories, Santa Monica, California), and the sensitivity of this assay was 0.22 ng/dL. Serum follicle-stimulating (FSH) levels were measured via enzyme immunoassays (FSH: Abbott Laboratories, Rome, Italy; LH: Dade Behring, Milan, Italy). The sensitivity of FSH assays was 0.2 mIU/mL.

Evaluation of reproductive development
All children underwent detailed examinations for reproductive development, which was scored by the same investigator referring to Tanner staging criteria based on development of genital, breast and pubic hair [16,17].

Phthalate metabolite measurements
Urinary phthalate metabolite levels were determined, as previously described [6]. Briefly, maternal urine was collected during the third trimester of pregnancy (28-36 weeks) and urine samples of children were obtained at age eight years. Urine samples were collected with a glass beaker. The samples were subsequently transferred into amber glass bottles and stored at 220uC for further analyses.  [18,19]. Metabolite concentrations were expressed as mg/L. Total DEHP levels indicated the sum of MEHP, 5OH-MEHP, and 5oxo-MEHP levels. Variability in urinary output was compensated for by adjusting phthalate metabolite concentrations to urinary creatinine levels. Metabolite levels that were adjusted for creatinine were expressed as micrograms per grams of creatinine (mg/gC). Urinary creatinine levels were measured at the Kaohsiung Medical University Chung-Ho Memorial Hospital using spectrophotometric methods, with picric acid as the reactive and the reader set at 520 nm.

Statistical analysis
Birth outcomes, children's characteristics, and sexual characteristics (Tanner scores) of girls were summarized as means and standard deviations (mean 6 SD) for continuous variables and n (%) for categorical variables. Sex differences were compared by using a two-sample t-test, if data were normally distributed, or a Mann-Whitney U-test, if data were not normally distributed, for continuous variables, and a Pearson Chi-square test for categorical variables. Maternal and offspring phthalate metabolites with adjustments for creatinine were summarized as geometric means and 95% confidence intervals (CI). Spearman's correlation analyses were performed to determine the relationships between maternal and offspring phthalate metabolite levels. Simple linear and binary logistic regression analyses were also performed to determine the factors associated with phthalate metabolite levels. Moreover, prior to conducting correlation and association analyses, birth outcomes, including birth weight, birth length, and head circumference, were adjusted for gestational age and expressed as a Z-score for analysis. All statistical assessments were two-tailed and considered statistically significant if P ,0.00625 (0.05/8). Statistical analyses were performed using SPSS 15.0 statistics software (SPSS Inc, Chicago, Illinois).

Participants' characteristics
A total of 130 children (61 boys and 69 girls) aged eight years, were enrolled into this prospective follow-up study. The birth outcomes, specifically birth weights, birth lengths, and head circumferences of newborns stratified by gestational age are shown in Table S1. Of the 130 children included in this study, only three had a gestational age of ,37 weeks (Table S1). Table 1 presents the birth outcomes, and children's physical characteristics and reproductive hormone levels at eight years of age of the entire cohort, as well as stratified by sex. At age eight years, girls had a higher BA and BA/CA ratio than boys (P,0.001). Prior to conducting correlation and regression analyses, the Z-scores of all birth outcomes for gestational age were calculated (Table S2). For the sexual characteristics of eight-year-old children, we did not observe any boy who had enter the Tanner stage 2. Of the 69 girls, 42 (60.9%) were classified at Tanner stage 1, 22 (31.9%) at Tanner stage 2 and 5 (7.3%) at Tanner stage 3 (Table 1). Table 2 presents the urinary phthalate metabolite levels (i.e., exposure levels) of mothers and their children at eight years of age. Boys had significantly higher urine creatinine levels than girls (48.7 vs. 27.71 mg/dL; P,0.001). After exposure levels were adjusted for creatinine, no significant differences were found between boys and girls in either in utero exposure levels and urinary phthalate metabolite levels at eight years of age.

Maternal urinary phthalate metabolite levels and birth outcomes
The correlations between birth outcomes and maternal urinary phthalate metabolite levels are presented in Table 3 and Table 4. No significant correlations were found between maternal urinary levels of phthalate metabolites and birth outcomes (Tables 3 and  4). Of the phthalate metabolites tested in this study, only in utero exposure to MMP had a negative correlation with birth weight in boys (r = 20.423, P = 0.006) (Table S3). In addition, urinary phthalate metabolite levels of children age eight years did not correlate with maternal urinary levels of phthalate metabolites (Tables 3, 4, Table S3). In addition, maternal urinary levels of phthalate metabolites did not correlate with the physical characteristics and reproductive development of children aged eight years (Table S4).

Phthalate metabolites of eight-year-old children and their development
The linear correlation between the physical characteristics and reproductive development of children aged eight years and their  Table 3. Effects of material urinary levels of phthalate metabolites on the birth outcome and the phthalate metabolites levels of eight-year-old children. urinary phthalate metabolite levels are presented in Table 5 and Table 6. Urinary MEHP levels correlated significantly with serum progesterone levels (b = 0.0004; SE = 0.0001; P = 0.006; R 2 = 0.060) (Table 5) (Table 6).
No correlations were shown between urinary phthalate metabolite levels of boys aged eight years with physical characteristics and reproductive development. Lastly, urinary phthalate metabolite levels of eight-year-old children did not correlate with either their physical characteristics or reproductive development (Table  S5).

Discussion
Epidemiological studies have indicated that phthalates may affect reproductive outcomes in children; however, the health effects of specific phthalates in children, in particular DEHP and its metabolites, are unknown [6,14]. In the present study, we examined birth outcomes, reproductive hormone levels, and markers of reproductive development of eight-year-old male and female children by following in utero and environmental exposure to phthalates. No significant differences were found in in utero phthalate exposure between boys and girls, indicating that both boys and girls in this study were exposed to similar levels of phthalate in utero.
Phthalates have been previously implicated in affecting fetal development, and in turn, birth outcomes. In a cohort of 404 infants, Wolff et al. [13] demonstrated that LMW phthalate metabolites, but not DEHP or HMW phthalate metabolites, were positively associated with gestational age and head circumference after adjusting for multiple factors, including creatinine levels. However, no associations were found between in utero phthalate levels and birth weights or birth lengths of newborns. In the present study, with the exception of one circumstance, in utero exposure to phthalates did not correlate with birth outcomes. Specifically, maternal urinary MMP levels were found to correlate negatively with birth weight in boys.
Phthalate metabolites have short biological half-lives and are rapidly excreted via urine [2]. This is supported by our finding that maternal phthalate levels did not correlate with those of their eight-year-old children. Therefore, we have originally comsidered that maternal/prenatal phthalate exposure and environmental phthalate exposure of 8-year-old children can be treated as two different entities that can be investigated separately. However, previous investigations have shown that prenatal phthalate exposure is associated with childhood behavior at ages 4-9 years [20], neurobehavioral development at ages 6-10 years [21], maletypical play (masculine play) behavior in boys at ages 3.6-6.4 years [22] and childhood eczema at age 2 years [23]. Therefore, as in the present study, it is also reasonable to hypothesize that prenatal phthalate exposure (maternal exposure) may affect sex steroid hormone levels and reproductive development of 8-year-old children. However, the outcomes of our study did not support the original hypothesis that maternal urinary levels of phthalate metabolites did not correlate with the physical characteristics and reproductive development of eight-year-old children (Table S4), which emphasizes the need for additional study.
Phthalates have been suspected to have anti-androgenic and estrogenic effects, which may alter the reproductive development of children [7,8]. In fact, some phthalates have been shown to alter or mimic E2 in vivo and in vitro [8]. In terms of environmental exposure to phthalates, as determined by the urinary phthalate metabolite levels of eight-year-old children, certain phthalate metabolites were indeed found to affect their physical characteristics and reproductive hormone levels. Regarding the reproductive hormone of the entire cohort, only MEHP correlated significantly with serum progesterone levels and MnBP correlated significantly with serum FSH levels. Urinary phthalate metabolite levels did not correlate with the physical characteristics and reproductive hormone levels of eight-year-old boys. Conversely, in girls, MEHP and MBzP were found to correlate positively with serum progesterone levels, and MnBP and MBzP correlated positively with serum FSH levels. Indeed, we found that for every 1 mg/L increase in urinary MBzP levels of girls aged eight years, there was an increase in serum FSH levels of 0.1726 mIU/mL (b = 0.1726, SE = 0.0471; P = 0.001). While exposure to some phthalates was associated with alterations in reproductive hormone levels in girls, none of the phthalates studied affected female reproductive development. These findings corroborate the results of another study, which demonstrated that phthalate exposure is not associated with precocious puberty in female children aged roughly seven years [24].
This study has a few limitations that need to be addressed. First, in addition to using a single urine specimen to determine maternal urinary phthalate levels during the third trimester of pregnancy, we did not measure phthalate levels during early pregnancy (i.e., the first trimester), which is when sex determination and differentiation of the fetus occurs. However, Braun et al. (2012) had previously determined that, with the exception of BPA and MBzP, the variability of most urinary phthalate metabolites was similar before and during pregnancy, suggesting that maternal urinary phthalate levels may be measured at any stage of pregnancy to determine in utero exposure to phthalates [25]. Second, given that we studied pre-pubertal children, it may be too early to determine the effects of phthalate exposure on gonadal development in these children by evaluating ovary follicles, endometrial development and uterus length. However, precocious puberty is defined as the development of pubertal changes at an age younger than the accepted lower limits for age at onset of puberty, namely, before age 8 years in girls and 9 years in boys [26]. Since we aimed to investigate the association between environmental phthalate exposure and abnormal reproductive development, we did not want to miss the disease onset time of CPP and thus started to follow the subjects at 8 to 9 years. Third, it was determined that urinary phthalate metabolite levels have an effect on certain sex hormone levels (i.e., FSH and progesterone in girls). For boys, sertoli cell markers (AMH and inhibin B) in serum are reliable markers for evaluation of basal testicular function in childhood [27], but we did not measure the serum levels of AMH and inhibin B in this research project. Of note, we have been continuously following this cohort of children since birth and they are now 12 years of age. Accordingly, we will investigate the effects of urinary phthalate metabolite levels on sexual development and sex hormone levels of 12-year-old children, including gonadal development of both boys and girls, and will measure the major markers of sexual maturation, including AMH and inhibin B.

Conclusion
In utero phthalate exposure did not significantly alter birth outcomes, growth, or reproductive function and development in pre-pubertal children. Furthermore, environmental exposure to certain phthalate metabolites (i.e., MEHP, MnBP, and MBzP) appears to affect reproductive hormone levels in pre-pubertal girls. Our results suggest that future studies focusing on the environmental health effects of phthalates and their metabolites are warranted.

Supporting Information
Table S1 Birth outcomes, including birth weights, birth lengths, and head circumferences, of newborns at different gestational ages. (DOC)

Author Contributions
Conceived and designed the experiments: SLW. Performed the experiments: PCL. Analyzed the data: HYC. Contributed reagents/materials/ analysis tools: PHS JYC CYL THY. Wrote the paper: PHS JYC. Specimen collections, tissue arrangement and preparations: HYC.