Endocrine-Disrupting Chemicals and Hippocampal Development: The Role of Estrogen and Androgen Signaling

Hormones are important regulators of key processes during fetal brain development. Thus, the developing brain is vulnerable to the action of chemicals that can interfere with endocrine signals. Epidemiological studies have pointed toward sexually dimorphic associations between neurodevelopmental outcomes, such as cognitive abilities, in children and prenatal exposure to endocrine-disrupting chemicals (EDCs). This points toward disruption of sex steroid signaling in the development of neural structures underlying cognitive functions, such as the hippocampus, an essential mediator of learning and memory processes. Indeed, during development, the hippocampus is subjected to the organizational effects of estrogens and androgens, which influence hippocampal cell proliferation, differentiation, dendritic growth, and synaptogenesis in the hippocampal fields of Cornu Ammonis and the dentate gyrus. These early organizational effects correlate with a sexual dimorphism in spatial cognition and are subject to exogenous chemical perturbations. This review summarizes the current knowledge about the organizational effects of estrogens and androgens on the developing hippocampus and the evidence for hippocampal-dependent learning and memory perturbations induced by developmental exposure to EDCs. We conclude that, while it is clear that sex hormone signaling plays a significant role during hippocampal development, a complete picture at the molecular and cellular levels would be needed to establish causative links between the endocrine modes of action exerted by EDCs and the adverse outcomes these chemicals can induce at the organism level.


Introduction
Throughout the second part of the 20th century, we have gained a deeper understanding of how environmental chemicals interact with and affect biological systems, including endocrine systems in vertebrates.Due to the fact that hormones control important aspects of organismal development and homeostasis, particular attention has been given to the study of chemicals that have the ability to interfere with endocrine functions and consequently cause adverse effects in the exposed organisms.Although the initial signal came from reproductive pathologies observed in farm animals and wildlife [1][2][3][4], the focus turned to humans after the discovery that a pharmaceutical estrogen called diethylstilbestrol caused an increase in the incidence of vaginal adenocarcinomas in girls exposed in utero [5,6].
Later on, the convergence of multiple streams of evidence, such as observations from wildlife, experimental laboratory studies, and human epidemiology studies, led to the creation of a coherent framework for understanding endocrine disrupting chemicals (EDCs).The World Health Organisation (WHO) published its first report on EDCs in 2002, defining these chemicals as "exogenous substances or mixtures capable of interacting with the endocrine system and leading to adverse effects in intact organisms, their progeny or (sub)populations" [7].In 2012, a second WHO report comprehensively documenting the available evidence on EDCs was published and concomitantly, the Organisation for Economic Cooperation and Development (OECD) released test guidelines for the assessment of EDCs (Guidance Document 150 on Standardised Test Guidelines for Evaluating Chemicals for Endocrine Disruption) [8,9].The OECD document recommends a tiered testing strategy ranging from in silico and in vitro studies to in vivo animal testing, albeit not all modes of action and adverse outcomes are covered.
Research into EDCs has continued to expand over time and increasing attention has been given to vulnerable windows of exposure and developmental effects of EDCs, including adverse outcomes on brain development [10].Because hormones are important regulators of key processes during fetal brain development, such as proliferation, apoptosis, differentiation, migration, and myelination (see, e.g., [11][12][13][14]), the developing brain is vulnerable to the action of chemicals that can interfere with endocrine signals [15,16].EDCs have been identified in biological fluids of pregnant women [17][18][19] and have been shown to reach the fetal brain [20].It is thus imperative to understand the potential consequences of neurodevelopmental exposures to EDCs and mitigate the potential risks through appropriate chemical testing and chemical safety regulations.
Epidemiological studies have pointed toward adverse cognitive and affective outcomes in children, linking prenatal exposure to EDCs with disorders indicative of developmental neurotoxicity (DNT), such as autism spectrum disorder, attention deficit hyperactivity disorder, and intellectual disabilities [21][22][23][24][25].These associative findings are supported by experimental data from laboratory animals, showing that prenatal/perinatal exposure to EDCs can lead to motor and behavioral abnormalities, as well as learning and memory deficits (reviewed in [26][27][28][29]).
Although the epidemiological findings are rarely stratified for sex, there are studies suggesting that some of the observed associations are sexually dimorphic [25,[30][31][32][33].
Several endpoints have been investigated for sex differences, including cognitive abilities in children, which can be differentially disrupted by EDCs in boys and girls [34][35][36][37].This is suggestive of interferences with sex steroid signaling during the development of neural structures underlying cognitive functions.We argue that this may be the case for the hippocampus, an old cortical structure of the limbic system, whose prominent role in learning and memory has been clearly established across taxa.During brain development, the hippocampus is subject to the organizational effects of estrogens and androgens, and thus it can represent a target for disruption by chemicals that can interfere with sex steroid signaling.In this review, we aim to briefly introduce hippocampus development, to then summarize the current knowledge about the organizational effects of estrogens and androgens on the developing hippocampus and present evidence that EDCs can perturb hippocampaldependent learning and memory functions.
Hippocampus: Structure, Function, and Development

Structure and Function
In humans, monkeys, and rats, the hippocampal formation is located in the temporal lobe, and its anatomical organization and functions are conserved across the three species [38][39][40].Structurally, the hippocampal formation is divided into areas including the cornu ammonis (CA), the dentate gyrus (DG), and the subiculum (Fig. 1).The CA is further divided into three regions called the CA fields, which are designated as CA1 to CA3.The CA fields contain pyramidal cells, whereas the DG contains granule cells.The subiculum is adjacent to the CA1 field and is connected to the entorhinal cortex (EC) on the parahippocampal gyrus of the temporal lobe [41].Within the hippocampal formation, the canonical route of information flow is represented through the trisynaptic circuit: the DG projects to the pyramidal neurons in CA3 via the mossy fiber pathway, the CA3 projects to the CA1 pyramidal neurons via the Schaffer collateral pathway, and the CA1 projects to the EC.The EC is not only a major efferent source, but also an afferent source, projecting to the DG (via the perforant pathway), the CA1, CA3, and the subiculum (reviewed in [42,43]).
Functionally, the CA fields and the DG can be further segregated along the dorsoventral axis (posterior-anterior in humans).While the dorsal hippocampus is primarily implicated in the processing of declarative memory functions, the ventral hippocampus is implicated in emotional responses (reviewed in [44,45]).The role of the ventral hippocampus in mediating emotional responses emerged from information on its functional connections with the amygdala, the hypothalamus and its regulatory control of the hypothalamic-pituitaryadrenal axis.With respect to hippocampal-dependent memory, evidence from genetic, lesion, and pharmacological manipulation studies shows that the dorsal hippocampus plays a prominent role in episodic memory, spatial learning, and contextual fear.Behavioral tests of spatial navigation have been particularly informative in this regard, and coincidentally, sex differences in spatial learning and memory are the most well-studied cognitive differences in rats due to the extensive use of spatial tasks in behavioral neuroscience [44,46].

Development
Hippocampal neurogenesis in rodents occurs from embryonic day (E) 10 to E18, and by E20 the hippocampus is largely formed [47].In humans, hippocampal neurogenesis occurs between gestational weeks (GWs) 16-18 [48].However, volumetric development persists to postnatal day (PND) 21 in rodents and to 2 years of age in humans, and this correlates to the maturation and integration of functional circuits [49,50].
The hippocampus develops in the dorsomedial region of the telencephalon.Early patterning of the hippocampus is orchestrated by the cortical hem, an important developmental organizer located near the telencephalic midline, between the emerging hippocampus and the choroid plexus (reviewed in [51]).The cortical hem provides morphogenetic signals including winglessrelated proteins and bone morphogenetic proteins to neighboring areas, and these are essential for the induction and correct organization of the adjacent hippocampus [51].Subsequently, hippocampal neurogenesis un-folds in spatiotemporal gradients which shape the patterns of connectivity and lead to the formation of neuronal circuits that underlie hippocampal functions [52,53].Although adult hippocampal function is dependent on experience-induced plasticity, it is apparent that correct sequential prewiring during development is also crucial for functional connectivity of hippocampal circuits (reviewed in [48]).
Through decades of research, we have gained an increasingly detailed understanding of neuronal ontogeny in the hippocampus and particular attention has been given to the fields comprising the trisynaptic circuit.In rodents, these start to develop around E11, with neuronal populations in each field arising in distinct, but overlapping gradients.Early autoradiographic studies in rodents show that the deepest layers of neurons in CA3 and CA1 are born first, with neurogenesis in these regions of Ammon's horn peaking at E14 and E15, respectively [47,52,54,55].DG neurogenesis peaks at E16, but unlike CA3 and CA1 neuronal populations, which are fully established before parturition, DG neurons continue to be born postnatally [47,55], and there is strong evidence suggesting that the DG retains its neurogenic potential into adulthood, which is important for learning and memory functions [56][57][58][59].However, a recent study in rats suggests that the early postnatal DG granule cells may have a unique function in hippocampal plasticity, possibly related to the weakening of existing memories in favor of learning and acquisition of new information [60].The protracted development of neurons in the DG is accompanied by a late maturation of granule cells, which persists beyond the first postnatal year in rhesus macaque monkeys and beyond the first decade in humans [61,62].The slow, postnatal maturation of hippocampal circuits is not unique to projections to and from the DG but also includes the CA fields [62].Overall, the morphological and functional maturation of hippocampal circuitry follows the pattern of afferent and efferent pathways and happens postnatally in rodents, primates, and humans.

Androgen and Estrogen Signaling during Hippocampal Development
Androgens and estrogens exert most of their cellular functions through binding to their respective nuclear receptors.Androgen receptor (AR), estrogen receptors alpha (ERα), and beta (ERβ) are nuclear transcription factors involved in the regulation of many complex physiological processes, including developmental processes relevant for brain development and brain sexual differentiation (reviewed in [63]).Although these signaling pathways have been extensively studied for their involvement in the organization of the sexually dimorphic hypothalamic nuclei that mediate reproductive behavior, it has become increasingly clear that they play important roles for the development of the hippocampus as well (reviewed in [46]).
In the developing human hippocampus from fetuses of unknown sex, the presence of ERα was revealed at GW 15 in CA pyramidal cells using immunostaining, and ERβ appeared 2 weeks later, at GW 17, in both CA pyramidal cells and DG granule cells.Around GW 25, ERα and ERβ reached hippocampal expression levels that were maintained into adulthood [64].Data on human fetal hippocampal AR expression and/or protein levels are lacking, but the receptors have been detected in the developing hippocampus of mammalian species such as rhesus monkey, sheep, and mouse [65][66][67][68][69]. Using in situ hybridization in the developing mouse embryo, Young et al. [67] detected AR mRNA expression in the hippocampus throughout E11-E18 in males, whereas females displayed significantly lower expression levels.Conversely, Mogi et al. [68] showed similar AR mRNA expression levels in hippocampi of both male and female mice embryos at E17 and E19.Postnatally, AR transcripts and protein levels increase dramatically, reaching similar levels in both sexes at PND21 [68,69].ER expression does not differ between sexes neither prenatally (E17, E19), nor postnatally (until PND7), and ER mRNA decreases rapidly to undetectable levels by PND7 [68,70].Hippocampal content of the major endogenous AR and ER ligands, testosterone (T), dihydrotestosterone (DHT), and 17β-estradiol (E2), peaks prenatally, but does not significantly differ between males and females in rats.After birth, levels of the three hormones decrease rapidly and without significant sex differences, albeit testosterone has a slower decline in males until PND8 [71].
The source of sex steroids is not only circulatory (reaching the brain from gonadal secretion), as the brain has been shown to synthesize steroids as well [72][73][74].Broadly, for T and E2, the steroidogenic process involves a multistep enzymatic conversion of cholesterol to T, which is then converted to E2 through the activity of aromatase [75].Local steroidogenesis has been reported for the developing hippocampus, particularly aromatization of T to E2, which has been shown to occur in limbic tissue homogenates (hippocampus and amygdala) of human, rhesus macaque, rabbit, and rat fetuses [76].Aromatase expression has also been detected in the hippocampus of fetal mice of both sexes, without significant differences in expression levels between males and females [77].Aromatase mRNA, protein, and enzymatic activity have been reported in the hippocampus of neonatal rats and mice [71,[77][78][79], and immunohistochemical staining confirmed its localization primarily in CA pyramidal neurons and DG granule neurons, with a weak presence in glial cells [79].Furthermore, mRNA expression of several important steroidogenic enzymes such as 3β-hydroxysteroid dehydrogenase, 17β-hydroxysteroid dehydrogenase, cytochrome P450 17α-hydroxylase, and 5α-reductase has been reported in the neonatal rat brain, together with their respective metabolic products, suggesting that biosynthetic pathways for hippocampal steroids extend beyond aromatization of circulating testosterone and potentially include complete de novo local neurosteroid synthesis [79].However, due to the difficulty of obtaining samples for analysis, data on steroidogenesis during prenatal human neurodevelopment in general, and in hippocampus in particular, is lacking.Although the presence of aromatase has been shown in cortical sections of second and thirdtrimester human fetuses, as well as in the cortices of newborns [80], future studies addressing the presence of steroidogenic enzymes in the fetal hippocampus would be useful for understanding hippocampal neurosteroidogenesis during human brain development.
Although the hippocampus is involved in spatial cognition, and the processing of spatial information is a sexually dimorphic function, the particular developmental contributions of androgens and estrogens to the observed sex differences in spatial ability have not been clearly delineated (reviewed in [46,81,82]).Human data show that men outperform women in multiple spatial tasks, and, similarly, rodent studies show that male rats consistently outperform females in maze tasks [83][84][85][86][87][88], supporting the hypothesis that spatial memory performance in mammals is sensitive to sex steroids.However, the extent to which sex differences in spatial cognition can be attributed to hormonally organized dimorphisms in neural structures is still difficult to ascertain.
In humans, there is evidence to support a developmental contribution of AR signaling to the organization of spatial abilities from subjects perinatally exposed to an excess of androgens or who lack functional ARs.For example, women with congenital adrenal hyperplasia show improved, male-like performance on the human analog of the Morris Water Maze and on Mental Rotations Test [89,90].Congenital adrenal hyperplasia is a genetic disease characterized by 21-hydroxylase deficiency and an inability to synthesize cortisol, which results in exposure to high levels of androgens beginning in gestation [91].Conversely, individuals with complete androgen insensitivity syndrome, caused by mutations in the AR gene that render the AR nonfunctional, perform worse than control men and women on the spatial subsets of Wechsler Intelligence Scale for adults [92].In addition, some evidence shows that estrogen deficiency in girls with Turner Syndrome (45, X) reduces their performance in spatial skills, while the evidence for boys with a testosterone deficiency due to Klinefelter Syndrome (47, XXY) is mixed [93][94][95].
Studies examining spatial ability in rats which lack functional ARs or have been prenatally treated with flutamide (AR antagonist) show that these animals exhibit poorer performance in the water maze as adults [96,97].Neonatal sex hormone manipulations in rodents show that male rats castrated or treated with cyproterone (AR antagonist) in the first 10 PNDs display inferior spatial learning as adults, compared to their intact or untreated counterparts [98,99], and this effect can be rescued by concomitant neonatal administration of testosterone in the gonadectomized animals [98].Interestingly, neonatal administration of testosterone, DHT, and estradiol to females masculinizes adult spatial memory performance [98][99][100], suggesting that both androgen and estrogen receptor signaling are involved in the early postnatal organizational effects on spatial memory.Furthermore, the organizational effects of sex steroids are associated with hormone-induced alterations in hippocampal morphology, which also correlate with spatial performance in adulthood.Compared to females, males have larger volumes of hippocampal DG granular cell layer, pyramidal cell layers in CA1 and CA3, as well as larger dendritic fields in CA3 neurons [96,98,100].This is suggestive of sex steroid signaling involvement in cell proliferation and/or apoptosis, dendritic morphology, and synaptogenesis in these regions.Evidence for ARinduced developmental effects shows that prenatal ex-posure to both testosterone and DHT promotes a larger field size in the DG and CA3 regions in females, possibly driven by dendritic arborization [96,98,100].Postnatally, testosterone also increases CA3 volume in females, whereas ovariectomy has no effect, suggesting that ER signaling is not essential for the morphological development of CA3 pyramidal neurons at this stage.However, in the absence of data on neonatal DHT administration, it cannot be excluded that testosterone may be aromatized into E2 locally in the hippocampus and that both AR and ER are synergistically involved in controlling CA3 dendritic growth postnatally.Neuronal differentiation and somal growth in the CA1 field seem to be sensitive to both AR and ER, as it has been shown that prenatal E2, but not DHT, can fully masculinize CA1 neuronal morphology in females, but the antagonist flutamide is able to feminize CA1 neuronal morphology in prenatally treated males [96,98].In addition, it has been shown that during the first postnatal week, males had substantially more proliferating cells in the hippocampus (DG and CA1) compared to females and that treatment with T, DHT and E2 significantly increased the number of proliferating cells in females [101].Interestingly, males also had more NeuN+ cells (differentiated neurons) and GFAP+ cells (differentiated glial cells) compared to females, but only T induced an increase in NeuN+ cells in females, whereas the number of GFAP+ cells was increased by both E2 and T [101].Overall, these data suggest that both steroids promote cell proliferation in the neonatal hippocampus, but T may promote neuronal lineage development, while E2 supports glial lineage development [101].
The role of sex hormones in contextual memory is less clear and has been mainly studied using various testing paradigms of fear conditioning in adult rodents and humans (reviewed in [102]).Although some sex differences have been reported, the organizational role of gonadal hormones in fear learning and its inhibition has not been studied.Further research in this area would be useful to understand why women are more prone to develop anxiety-related disorders and to establish chemical testing strategies able to discriminate sexually dimorphic adverse effects for anxiety-related phenotypes.
Lastly, sex differences in hippocampal function have also been studied from the perspective of object recognition and object location memory (reviewed in [46]).Although the involvement of the hippocampus in object recognition in rodents remains controversial [103,104], object location memory is a well-accepted hippocampalmediated function in both humans and rodents [105].However, it is important to note that the sexual dimorphism observed in object memory appears to be species Endocrine-Disrupting Chemicals and Hippocampal Development specific, whereby in humans, women exhibit enhanced object recognition and localization memory compared to men [106], but this difference is not readily apparent in rodents.To the contrary, male rats and mice show superior object localization memory compared to females, and studies on object recognition in rodents show conflicting results [46].Furthermore, the potential organizational effects of sex hormones in object memory have not been studied thus far and therefore, studies on neurodevelopmental effects of estrogen and androgen disrupting chemicals on object recognition should be interpreted with caution and in conjunction with the available supporting evidence.

Endocrine Disruption of Hippocampal Development
Studies examining the developmental impacts of EDCs on the brain have been emerging in the last decade, but there is still a scarcity of information related to less obvious targets of endocrine disruption such as the cortex, extrahypothalamic parts of the limbic system, the midbrain, and cerebellum [107].Given its fundamental role in mediating learning and memory functions, the hippocampus represents a particularly salient area of concern.
Since the hippocampus is developmentally organized by, and sensitive to, sex steroids, it is reasonable to suspect that exposure to EDCs that can interfere with estrogenic and androgenic signaling during brain development could alter the appropriate maturation of pathways essential for cognitive functions.Indeed, several EDCs such as bisphenol A (BPA), phthalates, triclosan, and certain pesticides have been shown to affect spatial cognition in rodents after developmental exposures (Table 1).
Bisphenol A BPA is a known EDC, used as a plasticizer in the manufacturing of polycarbonate and polyvinyl chloride plastics, among others [143].It has an ubiquitous presence in the environment and therefore gives rise to human exposure in various ways (e.g., through diet, household dust, and dermal absorption) [144].BPA has the ability to cross the placental barrier and is transferred into breast milk, thus reaching the developing fetus and infant [144].This EDC is recognized for its interference with ERs and the AR, acting as an estrogen, antiestrogen, and antiandrogen, depending on the cellular context.Because of this, the negative health consequences induced by BPA have been studied for decades (reviewed in [143]), including adverse effects caused by developmental exposures.
The topic of developmental BPA-induced adverse effects on brain development has also been investigated, including endpoints related to hippocampal-dependent learning and memory.The currently available literature regarding cognitive outcomes of developmental BPA exposure is highly heterogeneous in terms of study design, as the species, strain, dose, route, and duration of administration, as well as the type of testing and moment of testing differs across studies (Table 1).Although a straightforward comparison between these studies is not feasible, the main findings are summarized below, highlighting, where appropriate, important discrepancies.
In mouse, exposure to BPA through gavage of the pregnant dams from GD0 to PND23, followed by gavage of the offspring from PND24 to PND60, reduced spatial memory in the Morris Water Maze, concomitantly with a decrease in the excitatory neural circuits of CA3-CA1 and EC-CA1 [108].When the window of exposure is limited to gestation and lactation, and the exposure route is through maternal food intake, male performance was impaired in the Barnes maze, whereas females remained unaffected [109][110][111].A notable exception here is the California mouse strain, which did not exhibit any changes in spatial cognition in either sex, in the same developmental exposure and testing paradigm (gestation and lactation, through maternal food intake) [112].This is likely due to the fact that spatial cognition is not under strong sexual selection in the California mouse strain [145], and thus the positive control, ethinyl estradiol (EE), also failed to induce any changes in the Barnes maze performance of offspring of either sex [112].In rats, developmental exposure to BPA has also been shown to disrupt aspects of hippocampal-dependent memory in offspring, but there is not enough evidence to conclude whether or not these effects are sex-dependent.However, it is clear that compared to female mice, the developmentally BPA-exposed female rat offspring are more susceptible to cognitive impairments.Overall, BPA exposure throughout gestation and lactation has been shown to reduce rat offspring performance not only in maze tasks but also in an object recognition task and an inhibitory avoidance task [113][114][115][116][117]. In contrast to these results, there are three studies that have reported no developmental BPA effects on spatial cognition in the Morris Water Maze or the Radial arm maze [117][118][119].For two of these studies, the EE control is missing, calling into question the sensitivity of their experimental protocol to estrogen-mediated effects [118,119].The study by Ferguson et al. [117] has a well-designed protocol, which included naive controls (animals which were not Endocrine-Disrupting Chemicals and Hippocampal Development gavaged), EE controls, and a careful selection of a vehicle solvent devoid of estrogenicity.This study reported no effects of BPA or EE on offspring learning and memory in the Morris Water Maze and only an EE effect, in both sexes in the Barnes maze, where BPA did not significantly alter the latency to locate the escape box [117].Since the positive control EE did not elicit any effect in the MWM testing but did so at only one dose level in the Barnes Maze, the results of this study should be interpreted with caution, as it is not likely that estrogen-mediated BPA effects on hippocampal learning and memory would be detectable.Interestingly, the effect of the gavage procedure on the behavioral testing endpoints is evident, and it is possible that the stress induced by oral gavage is masking the effects induced by chemical exposure [146].
Generally, the BPA dosages employed in these rodent studies fall within one or two orders of magnitude of the tolerable daily intake (TDI) of 4 µg/kg/d set by the European Food Safety Authority, and in the majority of the cases are significantly below the developmental noobserved-adverse-effect level of 5 mg/kg for rats [147].Although human BPA exposure estimates indicate a low health concern [147], low-dose effects and nonmonotonic dose-responses (NMDRs) have not been evaluated for DNT endpoints.The study design and data analysis employed in the currently available in vivo literature on BPA do not allow conclusions on non-monotonicity for neurobehavioral endpoints, as either not enough doses are tested, there is no doseresponse, or the dose-response is not modelled in a way that would reveal an NMDR (see Table 1 for references).However, since several studies report effects at low doses (i.e., lower than 5 mg/kg) and BPA has been shown to induce putative non-monotonic effects on brain development [148], future well-designed studies should address this possibility and aid regulators in risk assessment.
At a cellular level, low-dose prenatal and perinatal BPA exposure in mice has been shown to disrupt hippocampal neuronal morphology and synapse formation, particularly in the CA1 region, where it induces a significant reduction in dendritic growth and spine density [120][121][122].This is suggestive of an antiestrogenic activity, as E2 has been shown to increase CA1 pyramidal cell dendritic spine synapse density in adult rats [149] and nonhuman primates [150], and BPA can antagonize the inductive effects of E2 on rat hippocampal dendritic spine synapse formation on pyramidal neurons in the CA1 [151].One possible mechanism through which BPA affects synaptogenesis could involve the regulation of N-methyl-D-aspartate (NMDA) receptors, which have been shown to be essential for mediating the effect of E2 on CA1 dendritic spine density [152].Indeed, low-dose developmental BPA exposure in rats has been reported to significantly decrease expression and protein levels of NMDA receptor subunits in the offspring hippocampus and this is accompanied by reduced expression and protein levels of synaptic proteins such as synapsin I, synaptophysin, spinophilin, and PSD-95 [122][123][124][125]. Since either mixed-sex pups or only male pups were used in these studies, we cannot conclude on whether the reported BPA effects on hippocampal synaptogenesis are sexually dimorphic.More recent studies show that developmental BPA induces sex-specific gene expression changes and epigenetic alterations in the rat hippocampus [126][127][128], although synaptogenic genes have not been examined.
EDCs have been shown to alter endocrine signaling by interference with hormone bioavailability [153], thus it is also possible that BPA affects the sexual differentiation of the hippocampus through indirect mechanisms, such as impairment in alpha-fetoprotein production.Alphafetoprotein is a liver-derived plasma glycoprotein that binds estrogens during prenatal life and it has been shown to be a key player in protecting the female mouse brain from masculinization and defeminization by estrogens [154].Indeed, low-dose developmental BPA exposure in mice can disrupt hepatocyte maturation and increases expression of alpha-fetoprotein in female, but not male offspring [155], and this could explain why prenatally exposed female mice fetuses do not show masculinization of spatial cognition.However, nonsteroidal estrogens such as BPA have a lower affinity for estrogen-binding glycoproteins compared to steroidal estrogens, and thus have a higher free blood fraction that is able to penetrate the cells and bind to nuclear ERs [156].Therefore, it is likely that low-dose developmental BPA can affect hippocampal-dependent learning and memory through both direct and indirect mechanisms, and the inclusion of plasma measurements for both BPA and sex-hormone binding proteins in future studies may help elucidate some of the inconsistencies in the previously reported results.

Phthalates
Phthalates represent a well-known class of EDCs, recognized for their antiandrogenic, androgenic, and estrogenic activity [157,158].These compounds are used as plasticizers, especially in the production of polyvinyl chloride plastics, but are also added to many other consumer goods.Similar to BPA, phthalates can leach out of materials and have become pervasive contaminants of the environment [159].Consequently, the widespread exposure of the population is reflected in the detection of phthalate metabolites in almost all human urine samples, with the major routes of exposure being food ingestion, personal care product use, and inhalation [160][161][162].
Exposure to phthalates has been associated with neurodevelopmental problems in children [22,[163][164][165][166] and impaired neurogenesis in vivo and in vitro [167,168].Although currently little is known about the effects of developmental phthalate exposure on learning and memory, there is emerging evidence to suggest that certain compounds of this class can sex-specifically impair spatial cognition after developmental exposure (Table 1).Di-(2ethylhexyl) phthalate (DEHP) has been shown to decrease spatial ability in young male, but not female, ICR mice exposed prenatally or throughout gestation and lactation [129,130].Concomitantly, the male mice had decreased levels of N-methyl-d-aspartic acid (NMDA) receptor subunits NR1 and NR2B in the hippocampus, suggesting an impairment in excitatory synaptic transmission [129].Other studies investigated molecular and morphological changes related to hippocampal function without linking them to learning and memory outcomes.In ICR mice, gestational and lactational DEHP exposure inhibited the phosphorylation of ERK1/2 in the hippocampus of pubertal mice and adult males, concomitant with reduced AR protein levels in pubertal hippocampus, and decreased ERβ expression in pubertal female and adult hippocampus of both sexes, without affecting circulating E2 or T levels [131].In male rats, developmental DEHP exposure induced a decrease in branching and total length of dendrites in hippocampal pyramidal neurons, whereas females remained unaffected [132].These morphological changes were accompanied by decreased protein levels of MAP2c, phosphorylated MAP2c, and phosphorylated stathmin, suggesting impaired dendritic development [132].In male, but not female rats, postnatal DEHP intraperitoneal injections reduced dendritic spine density and axonal innervation in CA3 neurons, decreased cell density in DG and CA3, and decreased hippocampal expression of Bdnf, a key neurotrophin involved in neuronal survival and growth [133,134].Brain-derived neurotrophic factor is a known target of E2 regulation [169,170], including during hippocampal development [171], and has been shown to mediate E2induced dendritic spine formation in hippocampal neurons in vitro and in vivo [172].In rat hippocampus, Bdnf is also regulated by E2-responsive miRNAs [173], and postnatal DEHP exposure has been shown to induce hippocampal miRNA downregulation in a dosedependent manner [135].Certain miRNAs, such as miR-132-3p, −212-3p, and −212-5p, which were down-regulated in males and upregulated in females, are part of the miR-132/212 cluster and are involved in memory formation and retention [174,175], and dendritic morphogenesis of hippocampal neurons [176,177].
Li et al. [136] investigated the neurobehavioral effects of dibutyl phthalate (DBP) after in utero and lactational exposure and found that spatial learning was enhanced in male rat pups at the highest DBP dose (675 mg/kg/day).Expression and protein levels of BDNF were concomitantly increased in the hippocampus of the high-dose DBP male rat pups [136].Conversely, another study found that developmental exposure to a slightly lower DBP dose (500 mg/kg/day) caused impairments in spatial learning only in male rat offspring and impairments in reference memory of immature rat offspring regardless of sex.These changes were accompanied by hippocampal neuron loss and synaptic dysfunction [137].While both DBP rat studies used the same rat strain and chemical exposure protocol (daily gavage of DBP dissolved in corn oil, from gestational day 6 to PND21), the animals were assessed in the Morris Water maze at different time points (PND22 vs. PND34 and 60) and beside measurement of escape latencies, which were evaluated in the same way, the authors performed and analyzed the probe trials with a hidden platform in different ways, which might explain the discrepancy in results.Indeed, it has been shown that developmental exposure to DBP 500 mg/kg/day can induce neurotoxicity in immature offspring rats through regulation of aromatase, ERβ, BDNF, and p-CREB protein levels, while it has no influence on mature offspring animals [138].Lastly, Boberg et al. [139] found that exposure to di-isononyl phthalate (DINP) from gestational day 7 to PND17 masculinized spatial learning in female rat offspring, suggesting a neurodevelopmental effect through antiandrogenic activity.
Overall, the available literature suggests that developmental exposure to phthalates can cause changes in hippocampal development and function, affecting learning and memory.Although not studied in this context, some of the observed neurodevelopmental outcomes are possibly mediated via ER-and/or AR signaling pathways.

Other EDCs
Triclosan (TCS) is an antimicrobial agent widely used in consumer products and is an EDC with estrogenic and androgenic activity [178,179].Due to its widespread use, TCS has been detected in the environment, as well as in human urine, blood, and milk samples [180][181][182].Although some in vivo data suggest that TCS is a developmental and reproductive toxicant [183][184][185][186], it is not Endocrine-Disrupting Chemicals and Hippocampal Development currently known whether long-term exposure during critical windows of development is associated with adverse neurobehavioral outcomes in humans.However, a recent study by Tran et al. [140] investigated neurobehavioral development of mice offspring after maternal TCS administration from E9.5 to PND28 and found that TCS impaired learning and recognition memory in the pups, without sex differences.Additionally, they showed that TCS impairs dendrite and axon growth in vitro and inhibits proliferation of neuronal progenitor cells, while promoting apoptosis [140].Further research would be needed to elucidate if these effects of TCS are at least partly mediated by effects on hippocampal development.
Polychlorinated biphenyls (PCBs) encompass over 200 different synthetic organochlorine congeners, which were used in various industrial and consumer products [187,188].Despite being banned several decades ago, PCBs persist in the environment due to their long half-life [189] and are still present in human adipose tissue and breast milk [190,191].While their deleterious effects on cognition are well-documented [192], underlying modes of action are not fully understood.A number of studies have demonstrated that PCBs disrupt thyroid hormone signaling, as well as dopamine neurotransmission and intracellular signaling [193].However, some PCBs, such as 2,2′,4,4′,5,5′-hexachlorobiphenyl (PCB 153), also display estrogenic properties [194,195] and have been shown to induce sexually dimorphic impairments on cognitive functions in rodents after developmental exposures [141,192].The extent to which these effects can be attributed to direct or indirect actions on sex steroid signaling in the brain remains to be addressed, but it is likely that complex interactions involving multiple hormone systems underlie the developmental CNS effects of PCBs [196].
Lastly, several compounds used as pesticides, belonging to a variety of chemical classes such as organophosphates, carbamates, organochlorines, and pyrethroids, have been shown to possess endocrine disrupting properties, affecting estrogen and/or androgen signaling in vitro and in vivo [197][198][199][200][201][202].Evidence for neurodevelopmental impairments is strongest for organophosphates, some of which are associated significantly and inversely with IQ scores in children (reviewed in [203]).Negative cognitive outcomes after gestational or postnatal exposures to organophosphates in experimental animal studies have also been reported, with a majority of studies using chlorpyrifos as a test compound [203].However, the effects reported on learning and memory tests are overall mixed and impossible to compare due to heterogeneity in choice of test, duration and route of treatment, age of testing, etc.For example, Gomez-Gimenez et al. [142] investigated the developmental effects of cypermethrin, endosulfan, carbaryl, and chlorpyrifos in rats after exposure from GD7 to PND21 and found that the outcomes on spatial learning were not only pesticide and sex-dependent but in some cases also depended on the type of test.Carbaryl did not induce any significant alterations in learning or memory in rat pups tested in the Morris water maze or Radial arm maze.Endosulfan and chlorpyrifos impaired spatial learning in the Morris water maze and Radial arm maze in males, but not in females.The most interesting finding was for cypermethrin, which improved learning in the Morris water maze in both sexes, but did not affect reference memory.Strikingly, cypermethrin impaired reference memory, but not working memory in the Radial arm maze in males, and improved working memory only in females.These discrepancies suggest that (at least partially) different neural pathways are involved in the two spatial tasks and underline the importance of test selection for the assessment of hippocampal-dependent learning and memory.

Conclusions
Although knowledge gaps still exist, it is clear that sex hormone signaling plays a significant role during hippocampal development at cellular and functional levels.Both ER and AR are involved in hippocampal neurogenesis in the perinatal period, and research to date suggests that both pathways influence key neurodevelopmental events such as cell proliferation, differentiation, dendritic growth, and synaptogenesis in the DG, CA1, and CA3 fields.These early organizational effects correlate with sexual dimorphism in spatial cognition and are subject to exogenous chemical perturbations.
Several EDCs that impact ER and/or AR signaling perinatally have been shown to affect spatial navigation in rodents, sometimes with sex differences.However, the extent to which the sexual dimorphic functional outcomes can be explained at the biological level is hindered in two significant ways.First, there are knowledge gaps regarding the specific developmental contributions of androgens and estrogens to the organizational processes particular to hippocampal development.A complete picture at the molecular and cellular levels would be needed to establish causative links between the endocrine modes of action and the functional outcomes.Second, there are differences between the testing paradigms used in the evaluation of chemically induced DNT.Standardized approaches in the choice of test, species, controls, route, and duration of administration would be useful to compare results from animal studies, and more in vivo data stratified by sex is needed.
Overall, research toward understanding sex hormonedependent mechanisms of brain development would help in addressing DNT induced by endocrine modes of action and cognitive impairments caused by developmental EDC exposures.This not only has scientific value but is also needed for improving testing and regulation of EDCs.While efforts are made to minimize exposure to chemicals with endocrine disruptive properties, the regulatory requirements for identifying a chemical as EDC include evidence that the substance has endocrine activity and that there are biologically plausible links between an endocrine mode of action and an adverse effect.Delineating the molecular and cellular underpinnings linked to disruption of cognitive functions mediated by the hippocampus and other brain regions will ensure the biological plausibility requirement for identification of EDCs.In addition, such research will also uncover key events that lead to adversity, toward which novel methods for chem-ical testing can be developed.Ultimately, this might contribute to reducing the burden of neurodevelopmental disorders that involve hippocampal function.

Fig. 1 .
Fig. 1.Organization of the hippocampus and the trisynaptic circuit.The entorhinal cortex (EC) connects to the dentate gyrus (DG) and the CA3 neurons via the perforant pathway (PP).The DG projects to the CA3 pyramidal neurons via the mossy fiber (MF) pathway and the CA3 projects to the CA1 pyramidal neurons via the schaffer collateral (SC) pathway.CA1 neurons send axons to the subiculum (Sb), as well as directly to the EC.

Table 1 .
Developmental EDC exposure studies in rodent models