Urinary mycoestrogens and age and height at menarche in New Jersey girls

Puberty is a critical developmental window during which there are rapid and significant changes in maturation and activation of multiple hormone axes [1, 2]. These changes make puberty a particularly susceptible time period for endocrine disruption. Exposures during this critical window can impact the physiology of female reproductive organs and the hypothalamic-pituitary-ovarian hormonal axis, including abnormal pubertal changes and ultimately, increased susceptibility to hormone-sensitive cancers including breast cancer [3].

Zearalenone (ZEN) is a mycotoxin produced by certain fungal species of the genus Fusarium, while alpha-zearalanol (zeranol, α-ZAL) is a synthetic derivative of ZEN as well as a product of ZEN metabolism. ZEN and α-ZAL are widely found in the food supply, where exposure may be deliberate (e.g. α-ZAL administered to livestock as a growth promoter) or inadvertent (e.g. fungal contamination of grain products by ZEN). Consequently, both ZEN and α-ZAL can be found through a range of food products including cereal, meat, milk, wine, beer, dried fruit and spices. Although these mycoestrogens have short half- lives (< 24 h) [4], continuous exposure through the diet presents a public health concern. In livestock, ZEN, α-ZAL, and their metabolites (i.e., α-zearalenol, β-zearalenol, zearalanol, β-zearalanol) are generally present due to the consumption of contaminated feed. These mycoestrogens have demonstrated endocrine disrupting effects in livestock and exposure can disrupt estrous cycling, resulting in decreased fertility [5, 6]. In fact, ZEN and zearanol are recognized as phytoestrogens, naturally occurring compounds that mimic the activity of estrogen, causing changes in hormone production and balance [7–9]. Despite widespread exposure to these compounds through diet [10–12], little is known about the effects of ZEN and α-ZAL exposures in human populations, particularly during critical windows of exposure such as puberty.

Several small studies, primarily in clinical populations, have measured ZEN exposure during the prepubertal period [13–15]. An Italian study (n = 63) examined 32 girls with central precocious puberty and 31 healthy controls [13]. Girls with detectable levels of ZEN (n = 6) and α-zearalenol were taller and proportionally heavier (p < 0.01) than girls with undetectable levels of these mycoestrogens. Asci et al. measured ZEN levels in girls living in the Mediterranean area of Turkey (n = 78) [14]. Girls were divided into three groups: girls with premature thelarche (PT), girls with idiopathic precocious puberty (IPP), and a control group. The authors reported that girls in the PT group had twice the levels of ZEN compared to the control group (p = 0.07). Similarly, girls in the IPP group had 2.8 times the levels of ZEN compared to the control group (p < 0.05). In the Jersey Girl study, a cross-sectional study of healthy 9–10 year old girls in the U.S., we previously reported that peripubertal girls with detectable ZEN in urine were of shorter stature (p < 0.0001) and less likely to have reached the onset of breast development than girls with undetectable ZEN (Prevalence ratio: 0.79, 95% confidence interval [95% CI]: 0.60, 1.04) [16]. We now extend these findings in the Jersey Girl Study cohort using follow-up data gathered for up to 10 years post-baseline. The aim of the current analysis is to examine baseline levels of ZEN, α-ZAL, and their phase-1 metabolites in relation to subsequent growth and onset of menarche. To our knowledge, this is the first longitudinal analysis of urinary mycoestrogens and pubertal outcomes in healthy girls.

This study is the first longitudinal study to evaluate ZEN exposure in relation to age at menarche and growth in healthy peri-pubertal girls. We observed that girls with detectable mycoestrogen levels were significantly shorter in stature at menarche than girls with mycoestrogen levels below the LOD. Girls with detectable α-ZAL concentrations were also less likely to have early menarche, but results were not statistically significant. The current study has expanded the findings from our previous report suggesting delayed pubertal development (i.e., thelarche) in girls with detectable mycoestrogen levels compared to girls with levels below the LOD [16].

Very few studies have measured mycoestrogen levels in relation to pubertal timing or growth [13–16]. Beyond our previous work in this population, a single study has evaluated body size [13] and no study has evaluated age at menarche. Our results contrast with those reported by Massart et al. suggesting that ZEN and α-zearalenol levels were positively associated with height and weight in girls. However, they measured mycotoxins in serum from girls (6–7 years old) participating in an idiopathic central precocious puberty case-control study. Additionally, they had a smaller sample size (n = 63) and only 6 girls had detectable levels of ZEN and α-ZAL. Finally, for those 6 girls with detectable serum ZEN levels, exposure (median ZEN: 0.12 ng/ml) was 10-fold lower than our measures in urine (median ZEN: 1.02 ng/ml). Thus given the differences in population, age at assessment, biospecimen type, and exposure levels, results of the two studies may not be readily comparable.

A larger literature has examined peri-pubertal growth and menarche in relation to other endocrine disruptors that act on sex steroid pathways. For example, in one study of adolescent girls, high intake of isoflavones, the predominant dietary phytoestrogens, was associated with older age at breast development [20]. Similarly in a diverse urban cohort, girls with higher levels of urinary phytoestrogens (i.e., daidzein, genistein, enterolactone) were less likely to have experienced Tanner Breast Stage 2+ than girls with low levels [21, 22]. Most recently, in the same cohort, Wolff et al. examined age at menarche in relation to urinary concentrations of phthalates and phenols, two classes of chemicals linked to altered estrogenic activity [23]. They showed that 2,5-Dichlorophenol was associated with early menarche (HR: 1.34; 95% CI: 1.06, 1.71) whereas enterolactone (HR: 0.82; 95% CI: 0.66, 1.03) and mono-3-carboxypropyl phthalate (HR: 0.73: 95% CI: 0.59, 0.91) were associated with later menarche. Finally, higher perfluorooctanoic acid (PFOA) and perfluorooctane sulfonate (PFOS) levels have been associated with delayed age of menarche [24]. Our results suggest that ZEN may have anti-estrogenic properties similar to those demonstrated in the phytoestrogen studies [20, 21, 23]. Like phytoestrogens, ZEN may inhibit the activity of key enzymes in the steroidogenic pathway including aromatase [25] and 17β-hydroxysteroid dehydrogenase [26], but more research is needed to fully understand these pathways.

Human exposure to ZEN is predominantly through food sources. ZEN can be found in the entire food chain through different products including cereal, milk, meat, dried fruit and spices. Although ZEN is not regulated in the U.S., the European Food Safety Authority (EFSA) established the tolerable daily intake for ZEN as 0.25 μg/kg body weight per day [12]. The European Commission also regulates limits for human consumption products (e.g., maize, cereals) and animal feed (e.g., maize, cereal, complementary feeding stuff). Our previous report suggested that beef and popcorn are the main sources of ZEN exposure in our study population [16], and meat (all types) was the primary source of ZEN exposure in local adult women [27], but additional work is needed to confirm these results and identify additional common sources of dietary exposure in humans.

ZEN metabolism plays a significant role in ZEN toxicity. ZEN follows two major biotransformation pathways: hydroxylation and conjugation [28]. Hydroxylation leads to α-zearalenol and β- zearalenol as well as α-ZAL, β-zearalanol, and zearalanone. Some of these hydroxylation metabolites tend to be more toxic than ZEN [29]. For example, in reporter gene assays examining estrogenic potency, ZEN was ~ 70 times less potent than α –zearalenol but twice as potent as β–zearalenol [29]. Conjugation results in ZEN glucuronide conjugates that are less toxic than ZEN. An in vitro study assessing ZEN biotransformation, reported that formation of ZEN glucuronides is a detoxification pathway that may attenuate estrogenic effects [30]. The proportion of ingested ZEN that goes through detoxification is not well understood and the relative contributions of the two biotransformation pathways may have different effects in human health. In the current study, ZEN urinary concentrations were higher than those of α-ZAL and other metabolites, suggesting that a portion of ZEN is excreted with no biotransformation. Other biomonitoring studies in adults have reported that some of the metabolites levels in urine are comparable or exceeded those of ZEN [31, 32]. We were not able to measure conjugated ZEN (or metabolites) which may explain the higher number of participant with undetectable levels of α-ZAL and other metabolites. Additionally, it may increase the potential for measurement error in our estimates. Exposure misclassification may be larger in the Cox regression results where exposure was categorized (<LOD, >LOD). It is difficult to speculate the effect of exposure misclassification in our results given the very limited toxicokinetics data on ZEN in humans. Furthermore, human biotransformation of these compounds may be influenced by sex, age, and diet adding to the variability of the biomarker. Additional biomarker(s) studies are needed to further investigate ZEN metabolism and toxicity.

It is well known that exposures during early life including childhood and puberty can influence cancer risk [1, 33]. For example, an estimated 10% reduction in breast cancer risk is expected for each two-year increase in age at menarche [34]. Growth metrics (including height and height velocity) during childhood have also been associated with cancer risk including breast, endometrial, testicular and prostate cancers [35–38]. Pubertal growth is mostly stimulated by estradiol and experimental studies have shown that estrogen may affect epiphyseal maturation, normal skeleton proportions, and bone mineralization [39–41]. Furthermore, there is some evidence that at low concentrations estrogens may accelerate linear growth whereas at high levels may stop linear growth [40]. ZEN is a structural estrogen analogue, able to mimic the activity of naturally occurring estrogens [9]. It has been hypothesized that ZEN can exert its estrogenic effect by affecting pituitary and gonadal function [9]. Experimental studies have shown that ZEN affects breast cancer development and progression through its estrogenicity [42–44]. In contrast with some of the experimental evidence, our results suggest that ZEN may also have anti-estrogenic properties. Differences in exposure levels (which are typically much higher in animal models than we observed in our study) and timing of exposure may also be important to consider. It is possible, for instance, that exposure to ZEN and/or its metabolites during critical developmental periods may suppress the production of 17β-hydroxysteroid dehydrogenase, ultimately resulting in anti-estrogenic properties.

Our study has several strengths and limitations of note. This is the first longitudinal study to evaluate the relationship between ZEN exposure and growth and pubertal outcomes in girls and we were able to retain 85% of participants through menarche. At the same time, like the other existing epidemiological studies of mycoestrogens, we had a relatively small sample size and exposure was based on a single urine sample collected at baseline. We were not able to measure conjugated metabolites which may lead to exposure misclassification and limit our ability to compare our results to other studies. While we had detailed anthropometrics at baseline, height and weight during the follow-up period was self-reported, which may have led to random error. However, this was reduced by the annual prospective data collection. Additionally, we had a homogenous population and differences in race and ethnicity could not be examined. This is important since some ethnic groups may be more vulnerable to mycotoxin exposures due to differences in dietary patterns [45], endogenous hormone concentrations [46, 47], and menarcheal timing [48]. In spite of these limitations, this study provides evidence that mycoestrogen exposure may impact girls’ growth and pubertal development. These results demonstrate a strong need for additional research to examine the impact of these chemicals on reproductive development at critical points in the life course.

Previous experimental evidence shows that mycoestrogens act as endocrine disruptors and their effect may depend on dose, hormonal environment, and critical windows of exposure. However, epidemiological studies examining important windows of exposure during childhood are lacking. Our study is the first one to examine mycoestrogens in relation to age at menarche. We demonstrated that girls with detectable mycoestrogen levels were significantly shorter in stature at menarche compared to girls with undetectable levels. Future studies in humans are extremely important given the evidence of widespread mycoestrogen exposure in human populations.