New Evidence that an Epigenetic Mechanism Mediates Testosterone-Dependent Brain Masculinization

Since the seminal publication of Raisman and Field (1) showing that neonatal exposure to testosterone organizes male-typical features of synaptic relationships in the rat hypothalamus, dozens of studies have reaffirmed the critical role of perinatal testosterone in the masculinization of aspects of brain morphology and function in rats and mice as well as in several other mammalian animal models (2). Several lines of investigation exploring the mechanisms of testosterone action in male-typical brain sexual differentiation have evolved since 1973 (3). Numerous studies have explored the contribution of perinatal testosterone to specifying the male-typical profile of neurotransmitter phenotypes in specific groups of neurons, of neuronal migration into specific sexually dimorphic nuclei, and in the expression of genes that control programmed cell death (apoptosis) within these nuclei. Other studies have examined the role of perinatal conversion of testosterone to estradiol (aromatization) and actions via estradiol receptors vs. direct actions of testosterone or its androgenic metabolite, dihydrotestosterone, on neural androgen receptors in the control of the above-named functions leading to male-typical brain sexual differentiation. Until recently, however, little attention has been paid to the molecular mechanisms whereby testosterone and/or estradiol modulate the expression of genes controlling the development of the male nervous system. A new study (4) published in this issue of Endocrinology points to a possible epigenetic action of neonatal testosterone in controlling the masculinization of the principal nucleus of the bed nucleus of the stria terminalis (BNSTp), in which neuronal number and nuclear volume are normally greater in male than in female mice (5). Epigenetic modification of chromatin structure due to the addition or deletion of acetyl groups to lysine residues of histone tails can lead to long-term changes in the expression of particular genes in the absence of any underlying change in DNA sequences. The critical observation of this new study was that neonatal administration of the anticonvulsant drug valproic acid (VPA), a histone deacetylase inhibitor, completely blocked the masculinizing action of either exogenous, neonatal testosterone in female mice or of testicular testosterone secreted neonatally in males on the development of BNSTp morphology.

In a preliminary experiment, Murray et al. (4) took pains to identify a dose (50 mg/kg) of VPA that reliably increased the levels of acetylated histone H3 protein in whole-brain homogenates within 24 h after its sc injection on the day of birth, without causing toxic effects. This outcome ensured that their dose of VPA effectively blocked endogenous histone deacetylase activity in the developing nervous system. The investigators then proceeded to administer this dose of VPA or saline vehicle on postnatal d 1 and 2 to groups of female mice that received either testosterone propionate (TP) or oil vehicle as well as to males that received oil vehicle on the day of birth. When brains were processed histologically at postnatal day 21, both cell number and total nucleus volume of the BNSTp were significantly greater in males and TP-treated females than in control females (all of these groups also received saline vehicle neonatally). By contrast, both TP-treated females as well as males that received VPA neonatally later had BNSTp cell number and nucleus volumes that were equivalent to those of females that received either saline or VPA neonatally. Importantly, Murray et al. (4) demonstrated considerable specificity of VPA treatment with their observation that neonatal administration of this histone deacetylase blocker had no effects on cell number or nuclear volume in either the suprachiasmatic nucleus or the anterodorsal thalamic nucleus (two brain regions that are not sexually dimorphic in mice). Additional evidence that the disruptive effect of VPA on testosterone-mediated masculinization of BNSTp morphology did not reflect a global teratogenic effect of this drug was the observation that there was no effect of neonatal VPA treatment on the development of the female BNSTp or on subjects’ body weights measured when all mice were killed on postnatal d 21. Finally, the authors argued persuasively that the effects of VPA did not result from an unwanted stimulation of GABA receptors in the brain insofar as such an action would have promoted hypothalamic masculinization (6), whereas in the present study, neonatal VPA blocked this process in the BNSTp.

The Murray et al. (4) results are among the first to implicate an epigenetic mechanism in any morphological aspect of brain sexual differentiation. These data provide an important complement to other recent reports pointing to epigenetic modulation of gene expression as an essential determinant of sexually dimorphic neural function in rodents. In one study (7), levels of acetylated histone H3 nuclear protein (H3K9/14Ac) in the cortex/hippocampus were significantly higher in male than in female mice killed at several perinatal ages, and transplacental treatment with testosterone augmented levels of H3K9/14Ac in females up to the level seen in males. No such sex differences or hormone effects were seen in either the hypothalamus or amygdala. Methyl-CpG-binding protein 2 (MeCP2) binds to methylated DNA where it recruits chromatin-binding factors, including histone deacetylases, that can either activate or repress gene transcription in the rodent forebrain (8). Kurian et al. (9) reported that MeCP2 expression was significantly greater in the amygdala and the ventromedial hypothalamic nucleus of female compared with male rats killed 1 d after birth. Subsequently (10), this same group reported that MeCP2 small interfering RNA, infused neonatally directly into the amygdala of both male and female rats, effectively suppressed MeCP2 protein levels in the amygdala and eliminated the sex difference in prepubertal play behavior in observations made over postnatal d 25–29. Thus, in the latter experiment, as in the current study of Murray et al. (4), the neonatal manipulation of epigenesis exerted long-lasting behavioral (10) or neural morphological (4) effects only in male subjects or in females given TP neonatally. Both results point to a possible role of epigenetic events in hormone-controlled behavioral and brain masculinization. It is well established that the masculinization of play behavior (11), like BNSTp morphology (4,12), depends on the neonatal actions of testosterone in the male rodent. It would be interesting to know whether neonatal infusions of MeCP2 small interfering RNA into the amygdala block the masculinizing actions of neonatal testosterone on amygdaloid morphology (13). In a similar vein, it would be important to know whether neonatal blockade of neural histone deacetylation in mice disrupts male-typical aspects of behavioral function that have been linked to the BNSTp. For example, BNSTp lesions selectively blocked the ability of female pheromones to elicit penile erections as well as mating behavior in male rats (14). Does selective neonatal inhibition of neural histone deacetylation in male rodents disrupt their ability to show either sexual arousal or mating behavior in adult tests with an estrous female? As pointed out by Murray et al. (4), lesions of the BNSTp were previously found to significantly augment activity of the hypothalamic-pituitary-adrenal (HPA) axis in male rats given restraint stress (15). Other previous research (16) showed that restraint stress more strongly activates the HPA axis in female than in male rats. Again, these data raise the question of whether neonatal inhibition of neural histone deacetylation would cause male rodents to show a female-typical profile of stress-induced activation of the HPA axis later in life. Murray et al. (4), like other authors (7,9) who have recently provided evidence of a contribution of epigenetic modulation of gene expression to brain and behavioral sexual differentiation in rats and mice, all point to the well-known sex differences in the incidence of autism spectrum disorder and attention deficit disorder (male > female) as well as bipolar disorder and Rhett’s syndrome (female > male) in humans. They argue persuasively that understanding the specific contributions of epigenetic mechanisms to brain and behavioral sexual differentiation in a variety of species, eventually including humans, may shed light on the etiology of the sex differences in the occurrence of these mental disorders.

Two specific experimental questions are raised by the results obtained by Murray et al. (4). The first is whether the disruptive effect of neonatal VPA on male-typical features of brain sexual differentiation extends to other well-studied brain regions that are sexually dimorphic in mice, including the anteroventral periventricular nucleus (AVPV) and the spinal nucleus of the bulbocavernosus. A second is whether promoters of particular genes implicated in male-typical brain sexual differentiation undergo epigenetic modification during perinatal periods when sex hormones are known to act. Forger and DeVries, the senior authors of the Murray et al. article (4), and their co-workers (5) previously showed that in mice, a null mutation of the Bax gene, whose protein product promotes apoptosis in the developing nervous system, abolished the sex differences in neuronal number in both the BNSTp (male > female) and in the AVPV (female > male). In future studies, it will be important to determine whether testosterone, or its neural metabolite estradiol, exerts epigenetic histone modifications of either the Bax gene promoter or of other neural genes that regulate programmed cell death in either the BNSTp or AVPV. Indeed, region-specific differences in the ability of testosterone to stimulate epigenetic modifications of cell death gene expression could help explain the paradoxical observation that endogenous neonatal testosterone augments neuronal number in the male rodent BNSTp while reducing neuronal number in the male AVPV in comparison with females that lack perinatal testosterone.