Sexual Behavior: From Hormonal Regulation to Endocrine Disruption

Sexual reproduction in vertebrates requires sexually differentiated behaviors to ensure attraction between partners and mating. These behaviors are a key factor in the success of sexual reproduction, on which species survival depends. Males and females adopt different behaviors and postures during attraction and mating. Male rodents readily express sexual behavior whenever a receptive female is present. During the precopulatory or appetitive phase, males engage in chemoinvestigation, displaying an olfactory preference for receptive females. Males generate ultrasonic vocalizations in response to females or their odors [1-3]. These vocalizations have characteristics in common with the songs of songbirds [4]. The emission of courtship vocalizations conveys information about the male’s motivational state and helps to attract the female partner. During the copulatory or consummatory phase, males mount, thrust, intromit. In male mice, mating ends with ejaculation, whereas male rats reach satiety after several ejaculations and do not then copulate again for 1–3 days [5].

In female rodents, receptivity is constrained by threshold levels of estradiol (E2) following progesterone secretion, and is, therefore, limited to the estrous phase of the cycle corresponding to the ovulation period. The female also participates actively in mating, through three phases of sexual behavior: attractivity, proceptivity, and receptivity [6]. During attractivity, the female stimulates male behavior by emitting pheromones. The female then adopts several proceptive behaviors (hops, darts, solicitations) in response to stimuli from the male [7]. Finally, during the copulatory phase, the female adopts a receptive posture called lordosis while approached from behind for insemination by the courting male [8].

Sexual behavior is tightly controlled by finely tuned neural processes, which begin during development and are tightly regulated by gonadal hormones. These processes may, therefore, be highly sensitive to exposure to endocrine disrupting compounds (EDC), defined by the WHO and the Endocrine Society as exogenous chemical substances or mixture of substances that alter the functions of the endocrine system. Exposure to EDC has been reported to alter several reproductive functions, including organ development, germ cell production, pubertal timing and other physiological processes required for fertility, as documented by both experimental and epidemiological studies [9]. Several studies have addressed the effects of exposure to EDC on mating behavior. Here, we review the experimental studies in rodents performed to address the potential effects of exposure to EDC on sexual behavior. We focus on the underlying molecular and neural processes affected by such exposure, paying particular interest to molecules with estrogenic and/or anti-androgenic activities. A recent review by Gore et al. [10] provided the general ethological background to sexual selection and reproductive competence in animal species and their sensitivity to EDC.

Since the pioneering work of Phoenix et al. [11], it has become clear that gonadal hormones play a key role in the sexual differentiation of mating behavior. This process begins early, in the perinatal (late gestational and early neonatal) period. In males, testosterone released from the fetal and neonatal testes permanently potentiates male (masculinization) behavioral and anatomic characteristics whilst inhibiting female (defeminization) characteristics in the neural circuitry underlying sexual behavior. This circuitry is stimulated by pheromonal cues emitted by receptive females and transmitted from the main olfactory epithelium and vomeronasal organ to the main and accessory olfactory bulbs, respectively, and then to the chemosensory responsive nuclei in the medial amygdala, the bed nucleus of the stria terminalis, and the medial preoptic area, where they are processed in behavioral responses. Projections are sent from the hypothalamic paraventricular nucleus to the spinal centers that promote penile erection and ejaculation, including the spinal nucleus of the bulbocavernosus, and the gastrin-releasing peptide system.

Perinatal testosterone secretion has organizational effects, resulting in structural, neurochemical and molecular differences in circuitry between the sexes. Differences in cell number and morphology or fiber density between the sexes have frequently been described for the medial amygdala, the bed nucleus of the stria terminalis, and the medial preoptic area [12]. For instance, the rat sexually dimorphic nucleus (SDN) and corresponding cluster of calbindin-immunoreactive neurons, both located in the medial preoptic area, contain more cells in males than in females [13, 14]. Conversely, neurons expressing kisspeptin and tyrosine hydroxylase in the anteroventral periventricular (AVPV) nucleus, a subdivision of the preoptic area involved in the ovulatory surge of LH, are more numerous in females than in males [15-17]. The regulation of these neuronal populations, or of other sexually dimorphic features, by perinatal testosterone can be mimicked by E2, because gonadal testosterone is aromatized into neural E2 by the aromatase cytochrome P450. At the molecular level, testosterone and its neural metabolite E2 also trigger differences in gene expression between the sexes, through long-lasting changes including epigenetic modifications [18].

The neural circuitry underlying behavior in females also includes the olfactory bulb, which transmits signals to the medial amygdala and then to the bed nucleus of the stria terminalis, the preoptic area, and the ventromedial hypothalamus, the principal facilitatory system for lordosis behavior. Systems inhibiting lordosis and involving the lateral septum, the preoptic area, and the arcuate nucleus have also been reported. These nuclei also receive projections from the medial amygdala and the bed nucleus of the stria terminalis. These inhibitory and facilitatory systems send projections to the periaqueductal gray regions through the ventromedial hypothalamus and lateral septum, respectively. These projections relay the information to spinal motoneurons innervating the axial muscles involved in the lordosis posture [19]. During the prenatal and early postnatal periods, these neural structures are protected from the masculinizing effects of sex steroids by the presence of α-fetoprotein, which selectively sequesters E2 derived from the mother and male siblings since the fetal ovaries are inactive [20].

The pubertal period is characterized by the central activation of pulsatile GnRH secretion, which stimulates the pituitary gonadotropin secretion required for sexual maturation and fertility. During this period, testosterone also exerts long-term effects on behavior to ensure the maturation of processes initiated during the perinatal period. Indeed, the prepubertal castration of male hamsters has been shown to decrease sexual behavior (fewer mountings and intromissions and longer times to ejaculation) and to shorten the time to lordosis relative to that for males castrated after puberty [21]. Neural processes, such as cell proliferation, in sexually dimorphic regions are influenced by gonadal hormones during the peripubertal period [22, 23].

In females, α-fetoprotein levels decrease after birth and the ovaries begin releasing E2 on postnatal day (PND) 7. Postnatal/prepubertal E2 secretion plays an active role in the feminization of sexual behavior. Indeed, E2 administration between PND15 and PND25 in aromatase knockout mice, which normally have highly impaired lordosis behavior, partially restores this behavior [24]. During this postnatal period, ovarian E2 is also essential for the establishment of the LH surge, which is synchronized with the receptivity period. In particular, prepubertal E2 activates an increase in kisspeptin expression in the rostral periventricular area of the third ventricle [25-27].

In adult males, gonadal testosterone acts on the male neural circuitry to stimulate sexual behavior. This activational effect of testosterone is transient by comparison to the permanent organizational changes induced during the developmental and pubertal periods. Male sexual stimulation is reduced or inhibited by castration but can be restored by hormonal supplementation. Testosterone and its neural metabolite E2 regulate the signaling pathways of neurotransmitters and neuropeptides playing an important role in displays of sexual behavior, such as oxytocin, dopamine, and glutamate.

Cyclic females mate only during the estrous phase and are sexually inactive during the rest of the cycle [28]. The preovulatory surge of E2, which occurs during the proestrus phase, triggers not only an ovulatory surge of LH, but also the expression of progesterone receptors (PR) in the ventromedial hypothalamus [29, 30]. Progesterone release under the control of LH then induces female receptivity, which is perfectly synchronized with ovulation in such species [28]. The increase in E2 levels also relieves constraints exerted by the inhibitory system through suppression of the inhibition exerted by the lateral septum and inactivation of the β-endorphin system in the preoptic area. Several neuropeptides and neurotransmitters present in the ventromedial hypothalamus display differences between the sexes. For example, estrogen receptors (ER), PR GABA, and enkephalin are all more abundant in females than in males and are known to promote female sexual behavior [31].

Sex steroids regulate sexual behavior principally through nuclear superfamily receptors. As the detailed mechanisms underlying the expression of sexual behavior have been largely reviewed, this paragraph summarizes very briefly the genetic studies investigating the relative contribution of androgen (AR) and ER. Indeed, the involvement of each of these receptors in this regulation has been studied by ubiquitous gene invalidations. In males, data from the testicular feminization mutation and global AR knockout models, both of which result in a feminine phenotype, have suggested that this receptor plays some kind of role in the expression of sexual behavior in rats and mice [32]. Global ERα knockout mice are infertile and have impaired sexual behavior [33]. Global ERβ knockout males have normal sexual behavior [34], but mutant males derived from an ERβ knockout mouse line completely devoid of ERβ transcripts [35] are infertile and display a mild impairment of sexual behavior [36]. In the global ERα and ERβ knockout models, females are infertile, with much lower frequencies of lor-dosis and proceptive behaviors than wild-type females [36, 37].

The neural role of these receptors was dissected more precisely by conditional gene invalidation, which has the advantage of preserving peripheral receptor expression and not interfering with gonadal and genital tract effects. In males, the estrogenic and androgenic pathways seem to play complementary roles in the organization and activation of the neural circuitry underlying sexual behavior. The perinatal masculinization and defeminization of brain structures involve the estrogenic pathway, whereas the neural AR is involved principally in the postnatal organization of spinal nuclei; however, both pathways seem to be involved in the adult activation of sexual behavior [32].

In females, neural ERβ invalidation delays the onset of puberty but has no effect on sexual behavior and fertility in either naïve or sexually experienced mutants [38]. This suggests that the ovary may be the primary site of ERβ action in these processes in adults. Further studies are required to address the behavioral phenotype of neural ERα knockout mice, but this receptor seems to play a major role in the E2-dependent regulation of reproductive processes, as demonstrated by the infertile phenotype induced by neural mutations [27, 39, 40].

We here describe the data collected for rodents up to December 2017, with a special focus on EDC exhibiting estrogenic/anti-estrogenic or anti-androgenic activities given the nature of signaling pathways involved in the expression of sexual behavior. Table 1 lists the EDC analyzed, with their estrogenic or anti-androgenic activities, uses and reference doses established by agencies when available. This list, which is far from exhaustive, includes the most widely studied molecules of these categories. Furthermore, some molecules, such as bisphenol A (BPA) or methoxychlor, are listed as estrogenic compounds, but can actually display anti-androgenic activities or target the thyroid system as for BPA. The keywords used to identify the studies concerned included the name of the molecule of interest (e.g., BPA, phytoestrogen, phthalate), and the terms “sexual behavior,” “endocrine disruptor,” “brain” or “nervous system.” Behavioral data are presented by period of exposure (developmental periods including prenatal, postnatal and pre/pubertal development, and adult exposure), type of activity of the compound (estrogenic or anti-androgenic), animal species (rat, mouse), and year of publication, in Tables 2 and 3, for males and females, respectively. When investigated in the same studies, neuroanatomical analyses of sexually dimorphic regions, neuropeptides or receptors involved in sexual behavior and hormonal measurements are included in Tables 2 and 3. Such analyses are otherwise presented separately in online supplementary Tables 1 and 2 (see www.karger.com/doi/10.1159/000494558 for all online suppl. material) for males and females, respectively.

E2, estradiol benzoate or analogs, such as ethinyl estradiol (EE) or diethylstilbestrol (DES), have been used as positive controls in studies evaluating compounds with potential estrogenic activity. The behavioral modifications observed differed between studies, ranging from behavior inhibition with blocked ejaculation for estradiol benzoate in males [41] to unaltered behavior for DES [42], as shown in Table 2. These discrepancies probably reflect the different doses of E2 and its analogs used. This, together with the use of only one dose of the positive control in most studies, makes it difficult to draw any firm conclusions about the potential estrogenic effects of the EDC molecules studied. We will therefore discuss the reported effects of exposure to BPA, phytoestrogens, nonylphenol or methoxychlor without comparison to positive controls.

Prenatal/Postnatal Exposure. Twelve studies in rats and 7 in mice have investigated the effects of early exposure to estrogenic compounds, such as BPA, phytoestrogens (coumestrol, genistein, resveratrol) nonylphenol and methoxychlor (Table 2).

In rats, with the exception of high doses of NP, which did not affect male sexual behavior [43], changes in at least one of the components of sexual behavior were reported. Increases in the time to intromission and a larger number of intromissions were reported for prenatal and postnatal exposure to BPA, respectively [44]. A dose of 100 µg/kg/day of BPA decreased the intromission ratio without modifying the number of mounts, mating duration or time to ejaculation in rats [42], and a decrease in time to intromission was associated with no change in the number of intromissions or the intromission ratio [45]. One interesting study reported different behavioral responses as a function of BPA dose and sexual experience [46]. Males exposed to phytoestrogens were reported to have a lower frequency of mounting, intromission and ejaculation in a number of studies [42, 47-50], whereas 2 studies reported increases in these behaviors [41, 51]. In middle-aged males, chronic exposure to methoxychlor has also been shown to decrease time to ejaculation [52]. The published neuroanatomical data are presented in online supplementary Table 1: 6 studies described changes in the SDN or AVPV volume or cell number [47, 53-57], whereas 3 studies reported no effect [42, 58, 59]. In the studies reporting effects of exposure on the organization of sexually dimorphic populations, both similarities and differences were observed, depending on the type of neuronal cell analyzed. Indeed, increases in the number of tyrosine hydroxylase cells or AVPV volume have been described [47, 56, 57], suggesting demasculinization and/or feminization of this hypothalamic region. Opposite patterns have been observed, with either a decrease (resveratrol [47]; BPA [54]) or an increase in SDN volume or calbindin cell number (genistein or nonylphenol [55]; BPA or genistein [57]; BPA [53]). Overall, such neuroanatomical measures alone are probably not sufficient or enough sensitive to reflect EDC-induced effects on the expression of behavior.

In mice, males exposed prenatally to methoxychlor displayed diminished sexual arousal, spending less time close to a female partition [60]. Exposure to BPA or phytoestrogen did not affect olfactory preference, time to the initiation of certain behaviors or the frequency of behavioral events [61-64]. Decatanzaro et al. [61] observed only a fewer number of intromissions in animals exposed to BPA in a high-phytoestrogen diet. In addition, male mice exposed to BPA displayed less lordosis behavior than the vehicle group when gonadectomized and primed with E2 and progesterone, indicating a greater masculinization of behavior [63]. When given a choice between control and BPA-exposed partners, female mice preferred controls over BPA-exposed males [65]. BPA had no neuroanatomical effects on the number of calbindin-immunoreactive neurons in the preoptic area or kisspeptin cells in the AVPV [63].

Thus, exposure to BPA or phytoestrogens induces changes in the neural circuitry underlying sexual behavior in rats. Little if any change was observed in mice, suggesting species differences, although more studies are required to confirm these observations. One key observation is that most of these studies were performed at relatively low doses, equivalent to or lower than the reference doses established for these molecules.

Adult Exposure. Exposure to BPA for 2 weeks has no effect on male rats [66]. By contrast, longer periods of exposure of male mice to BPA increased the times to mounting, intromission and ejaculation, and decreased the numbers of intromissions and thrusts in both naïve and sexually experienced male mice [63]. Similar differences between short- and long-term exposures were observed for corncob bedding and derived tetrahydrofuran-diols [67, 68]. Changes in behavior were also reported for rats subjected to long-term phytoestrogen exposure [69] and NP [70]. A non-monotonic dose response was observed for NP since increased emission of ultrasonic vocalizations, reduced number of mounts, intromissions and thrust as well as delayed ejaculation were observed only at the dose of 5 µg/kg/day but not at lower and higher doses. Despite the small number of studies performed, the results obtained all suggest that long-term exposure in adults affects male sexual behavior.

Prenatal/Postnatal Exposure. Female sexual behavior has mostly been investigated in rats (13 studies vs. only 3 in mice). BPA exposure had no significant effect on lordosis quotient or proceptive behavior in a number of studies [42, 44, 46, 71-74]. However, pre- or postnatal exposure to BPA decreased proceptive behavior in one study [75] and increased the frequency of lordosis in another [44]. In one study on mice, exposure to BPA at a dose of 50 µg/kg/day resulted in a significant increase in lordosis quotient in naïve but not sexually experienced females, whereas a dose 100 times higher had no effect [76]. Olfactory preferences were unchanged, as BPA-exposed females displayed normal levels of interest in chemoinvestigating males relative to females [76]. Following ovariectomy and testosterone treatment, females displayed no masculinization of their behavior, with no difference in mounting or thrusting behaviors relative to the vehicle-treated group. Most neuroanatomical studies have reported no effects on the SDN, corresponding calbindin cells [42, 53, 54, 59, 73, 76], and TH neurons [56]. By contrast, an increase in the number of kisspeptin neurons in the AVPV was observed in the study reporting an increase in lordosis behavior following BPA exposure in mice [76]. Changes in the number of tyrosine hydroxylase neurons have also been reported in rats [54, 56]. These data strongly suggest that BPA does not trigger masculinization in the female neural circuitry during the highly sensitive perinatal period, despite its lack of binding to α-fetoprotein. Instead, it seems to potentiate estrogen-induced increases in sexual behavior and kisspeptin expression, both of which occur later during the postnatal/prepubertal period in females.

By contrast to BPA, prenatal or postnatal exposure to phytoestrogens mostly affected adult female behavior, in both rats and mice, as shown in Table 3 [47, 51, 62, 77-79]. Similar results have been reported for methoxychlor [80]. More behavioral and neuroanatomical analyses in the same context would be required to determine whether the observed alterations are linked to changes in the organization of sexually dimorphic populations. Indeed, no effects of resveratrol on SDN-preoptic area or AVPV volume were found in a study on female rats in which no change in behavior was detected [47]. Two other studies reporting an absence of effect on these two hypothalamic regions did not assess female behaviors [56, 64].

Adult Exposure. The acute exposure of ovariectomized rats to BPA did not trigger lordosis behavior when females were primed with progesterone only, or with progesterone plus E2, despite an increase in the number of progesterone-immunoreactive cells in the preoptic area and ventromedial hypothalamus [81]. An analysis of 10 sets of findings for adult exposure to phytoestrogens suggested that behavioral responses differ between initial hormonal environments. Hence, genistein and daidzein, and ferutinin administered to ovariectomized females primed only with progesterone either induced [82-84] or had no effect on female behaviors [85, 86]. In similar experimental conditions, DES treatment induced lordosis behavior [87], whereas the anti-estrogen tamoxifen had no effect [86]. By contrast, as for tamoxifen [86], exposure to genistein and daidzein, corncob bedding, ferutinin, resveratrol or Humulus lupulus reduced lordosis behavior in 5 studies in which females were either intact or ovariectomized and on E2 supplementation [68, 82, 83, 86, 88]. In 3 other studies, however, genistein, resveratrol and H. lupulus had no effect on lordosis behavior but altered other components of behavior, such as the frequency of anogenital investigation or proceptive behavior [85, 89, 90]. In the absence of E2, exposure to methoxychlor also induced proceptive and lordosis behavior in females after progesterone injection [91]. These data suggest that phytoestrogens and methoxychlor have estrogenic activity in the absence of endogenous estrogens, but anti-estrogenic effects in the presence of endogenous or supplemented estrogens, particularly for phytoestrogens.

Five of the 6 studies addressing the developmental effects of vinclozolin and phthalates at relatively high doses on components of sexual behavior in males reported behavioral changes [92-96]. Diminished erections and increased seminal emission were observed, together with reduced attractiveness, in rats exposed to vinclozolin [92, 96]. Exposure to phthalates led to sexual inactivity [95], increases in the time to sexual behaviors and a decrease in the frequency of these behaviors in rats [93, 94]. Only the study by Andrade et al. [97] reported an absence of effects for low or high doses of DEHP. One neuroanatomical study reported no effect on SDN-preoptic area volume of exposure to high doses of DINP (diisononyl phthalate) [58]. One recent study addressed the effects of adult exposure to low doses of DEHP on male sexual behavior. Such exposure decreased ultrasonic vocalizations, reduced female attraction and delayed the initiation of mating, without affecting olfactory preference for receptive females [98].

Only one study has measured the effects of developmental exposure to phthalates on female sexual behavior. It reported a decrease in the lordosis behavior of adult females in proestrus following exposure to DBP (dibutyl phthalate), DINP, and DEHA (di-[2-ethylhexyl] adipate) [94]. The effects of adult exposure to anti-androgenic compounds have not been studied.

More studies of developmental and adult exposure to anti-androgenic molecules are therefore required, in both males and females, particularly for doses close to the estimated environmental exposure.

Endocrine disruption of sexual behavior may occur through indirect or direct pathways (Fig. 1). The indirect pathway involves changes in the levels of gonadal hormones, which then affect the organization or activation of neural structures involved in the expression of behavior. The circulating levels of gonadal hormones may be modified by dysregulation of the hypothalamic GnRH system and upstream regulators including kisspeptin neurons and/or the disruption of pituitary function or gonadal steroidogenesis. Exposure to EDC can also directly affect the neural structures underlying sexual behavior, by interfering with the neural synthesis of hormones, such as the aromatization of testosterone into E2, binding to and activation of sex steroid receptors, or the expression of these receptors. Modes of action involving agonism or antagonism are more difficult to demonstrate in vivo. They can be suggested or supported by parallel comparisons with the effects of positive controls, such as estrogens or their analogs, or anti-androgens, such as flutamide, although such comparisons may be limited for the reasons described in the section “Effects of Developmental versus Adult Exposure to Estrogenic Compounds”. Analyses of the modes of action of various compounds are reported in Tables 2 and 3 when performed in parallel to behavioral studies; otherwise, they are presented separately in online supplementary Tables 1 and 2.

Three of 12 studies on male rats reporting developmental effects of EDC on behaviors described a decrease in adult hormone levels following exposure to BPA [45], genistein [50], and resveratrol [47]. The other 9 studies found no effects on adult levels of testosterone, E2, and gonadotropins [42, 43, 48, 49, 52, 93, 94]. In mice, no change in hormone levels was observed following exposure to BPA or EE [61, 63, 99]. Exposure to methoxychlor did not change basal hormone levels but prevented the increase in testosterone levels following exposure to females [60]. The observed impairment of male sexual behavior does not, therefore, seem to be systematically linked to changes in hormone levels. However, these findings do not rule out the possibility of transient changes in hormone levels during the exposure period, particularly if exposure occurs during development, thereby inducing long-term changes to the neural processes or hormonal signaling pathways underlying male behavior. Interestingly, long-term changes in the levels of the sex steroid receptors, ER and PR, have been reported to be induced by prenatal and postnatal exposure to BPA, DES or methoxychlor in the whole hypothalamus or more specifically hypothalamic areas (preoptic area, ventromedial hypothalamus), bed nucleus of stria terminalis or medial amygdala of rats [100-102], and mice [103, 104]. However, no effects of such exposure were reported in 2 other mouse studies in the same brain areas [105, 106]. Changes in total hypothalamic gene expression were observed for BPA and EE, or for neuropeptides involved in social behavior, such as AVP in the medial amygdala and lateral septum and oxytocin in the hypothalamus [105, 107].

In females, sexual behavior is generally analyzed in ovariectomized animals supplemented with E2 and progesterone. The behavioral changes observed under these experimental conditions cannot be due to modifications of the hypothalamic-pituitary-gonadal axis. However, it is important to analyze hormone levels in intact females in parallel, because they can provide information about the integrity of the hypothalamic-pituitary-gonadal axis and its effects on the activation of behavior under physiological conditions. It is not possible to draw any firm conclusions from the small number of studies assessing hormone levels in intact females. An absence of effect [42, 105, 108] or increases in hormone levels have been reported following exposure to BPA [76]. Lower levels of estradiol and LH were observed in rats exposed to methoxychlor [109] or BPA [110], whereas phthalates had no effect [94]. By contrast, changes in sex steroid receptor levels have been reported in a larger number of studies. Most of the studies on BPA presented in online supplementary Table 2 dealt with the effects of exposure on ERα and/or ERβ levels in neural structures underlying neuroendocrine and behavioral reproductive functions such as the medial amygdala, the bed nucleus of stria terminalis, the preoptic area and its subdivisions, the ventromedial hypothalamus or the arcuate nucleus in rats and mice [100, 101, 103, 105-109, 111-113]. This was probably motivated by the facts that these receptors are the major mediators of estrogen-induced genomic regulation, and effects of EDC exposure on their neural levels can interfere with of the expression of reproduction behaviors. The differences observed in the effects induced may reflect differences in the doses of BPA used, the period of exposure, and the area of the brain analyzed. The few studies to have assessed ER expression in parallel to behaviors reported lower levels of ERα expression in the preoptic and ventromedial nuclei of females with low levels of proceptive behavior [75], or no change in the number of ERα-immunoreactive neurons in the medial amygdala, bed nucleus of stria terminalis and preoptic area of animals with normal or higher levels of proceptive behavior [72, 76]. Exposure to methoxychlor has also been shown to trigger changes in ER and PR levels in the preoptic area [102, 110]. Some studies pointed out the involvement of long-lasting epigenetic modifications in such changes. Epigenetic gene regulation includes DNA methylation at CpG sites, post-translational histone modifications such as methylation or acetylation, and microRNAs. A previous study showed that changes induced by BPA exposure in the expression levels of ERα in the hypothalamus are associated with modifications in DNA methylation in the promoter region of this gene [103]. Similar observations were made in the preoptic area of females exposed to methoxychlor [110]. These findings are of particular interest given the role of ER in the expression of female sexual behavior. Only a few studies have monitored the effects of EDC exposure on other targets, such as neuropeptides essential for the expression of reproductive behaviors [107, 114].

In male rats, changes in male sexual behavior following exposure to phytoestrogens were associated with lower levels of testosterone [69]. In mice, the impairment of sexual behavior following chronic adult exposure to BPA, NP or DEHP was associated with unchanged hormonal levels and integrity of the hypothalamic-pituitary-gonadal axis, together with changes in the level of sex steroid receptor expression [63, 70, 98]. In particular, DEHP down-regulated AR protein and mRNA levels in the neural circuitry involved in sexual behavior [98], while both the numbers of AR and ERα-immunoreactive cells were affected by NP exposure [70].

The exposure of adult females to resveratrol or BPA increased E2 levels in both rats and mice [85, 115]. Interestingly, regardless of its behavioral effects, exposure to phytoestrogens or BPA increased the levels of ERα, ERβ or PR expression in the preoptic area, the ventromedial hypothalamus or the paraventricular nucleus [81, 83, 85, 88, 90, 116].

Several observations can be made based on the data concerning the impact of EDC on sexual behavior reviewed here. In general, sexual behavior appears to be highly sensitive to EDC, and can be added to the endpoints generally used in assessments of the risks of exposure to EDC and their potential impact on reproduction. Like other reproductive endpoints, the behavioral effects induced by exposure to EDC depend on the period of exposure. The developmental and pubertal stages are particularly vulnerable due to the organizational effects of hormones during these periods, but exposure in adults may also have effects, the doses used in several studies having elicited effects at doses below the reference dose. Effects may also depend on the sex of the animals.

The amounts of data published differ considerably between EDC compounds. The most frequently investigated compounds are estrogen-like molecules, such as BPA and phytoestrogens. However, published studies have not addressed all possible periods of exposure in both sexes. Developmental exposure to BPA or phytoestrogens has been extensively studied in both males and females, but adult exposure to phytoestrogens has been studied only in females. Much more studies are required for other estrogenic or anti-androgenic compounds, for which fewer data are available, regardless of the exposure period considered.

Another interesting observation is that changes in the level of expression of sex steroid receptors (ER, AR, PR) have been widely documented in both males and females. This finding strongly suggests that EDC may act, at least partly, through changes in the neural signaling pathways underlying sexual behavior, either directly through epigenetic modifications to the ERα promoter, as reported for developmental exposure to BPA [103], or through as yet unidentified pathways, as for the down-regulation of AR induced by the exposure of adults to DEHP [98].

Finally, although caution is required when extrapolating findings from rodents to other species, these altered processes may be of considerable relevance in both humans and wildlife. The regulation of libido and erectile function or reproductive behaviors by sex steroid hormones at the neural level is highly conserved across species. In this context, human studies have reported an association between a decrease in sexual activity and exposure to environmental doses of DEHP and BPA [117, 118], but the mechanisms underlying these effects remain unclear.