Sex-specific behaviors occur because the nervous system differs between males and females. These differences arise from a combination of genetic factors, such as components of the sex determination pathways, as well as environmental factors, such as social experience. In many cases both genetic and environmental inputs act through the steroid hormone system to sculpt the nervous system. Many anatomical instances of sexual dimorphism have been documented, including differences in the numbers and size of neurons in particular structures as well as differences in the pattern and number of synapses.
It is challenging to trace the chain of causality from environmental or genetic factors to the development of neural dimorphisms and to link these differences to sex-specific behaviors. In this section we examine a few cases in which studies in experimental animals have provided insights. In later sections we ask whether similar mechanisms underlie sexually dimorphic behaviors in humans.
However, before proceeding we note that the ways in which chromosomal mechanisms of sex determination are linked to the cellular processes of sexual differentiation in the central nervous system vary widely among species. In insects sex differences in behavior are independent of hormonal secretion from the gonads, and instead rely exclusively on a sex determination pathway within individual neurons. This mode of sexual differentiation of the brain and behavior is particularly well understood in the fruit fly, where it has been demonstrated that the sex determination cascade initiates expression of a single transcription factor, fruitless, that specifies the entire repertory of male sexual behaviors (Box 58–1).
Box 58–1 Genetic and Neural Control of Mating Behavior in the Fruit Fly
In the presence of a female fruitfly the adult male fly engages in a series of essentially stereotyped routines that usually culminate in copulation (Figure 58–6A). This elaborate male courtship ritual is encoded by a cascade of gene transcription within the brain and peripheral sensory organs that masculinizes the underlying neural circuitry.
Sex determination in the fly does not depend on gonadal hormones, as it does in vertebrates. Instead, it occurs cell autonomously throughout the body. In other words, sexual differentiation of the brain and the rest of the body is independent of gonadal sex. The male-specific Y chromosome of fruit flies does not bear a sex-determining locus. Instead, sex is determined by the ratio of X chromosome number to autosome number (X:A). A ratio of 1 is determinative for female differentiation, whereas a ratio of 0.5 drives male differentiation.
The X:A ratio sets into motion a cascade of gene transcription and alternative splicing programs that leads to the expression of sex-specific splice forms of two genes, doublesex (dsx) and fruitless (fru). The dsx gene encodes a transcription factor that is essential for sexual differentiation of the nervous system and the rest of the body, with the sex-specific splice variants responsible for male- and female-typical development.
The fru gene encodes a set of putative transcription factors that is generated by multiple promoters and alternative splicing. In males one particular mRNA (fruM) is translated into functional proteins. In female flies alternative splicing results in the absence of such proteins.
Males carrying a genetically modified fru allele that can only be spliced in the female-specific manner (fruF) have essentially normal, dsx dependent sexual differentiation. These fruF males therefore resemble wild type males externally. However, the loss of FruM in these animals abolishes male courtship behavior directed toward females. These data indicate that FruM is required for male courtship and copulation.
Conversely, transgenic female flies carrying a fruM allele exhibit male mating behavior toward wild type females, indicating that fruM is sufficient to inhibit female sexual responses and promote male mating.
Intriguingly, fruF males do not court females and, like wild type females, do not reject mating attempts by wild type males or fruM females. Similarly, fruM females attempt to mate with both fruM and wild type females. These data suggest that fruM may also specify sexual partner preference, which in the case of wild type males would be directed to females.
In wild type females without fruM the neural pathways are wired such that these flies exhibit sexually receptive behaviors toward males. When groups of fruF males (or fruM females) are housed together, they court each other vigorously, often forming long chains of flies attempting copulation.
To build the circuitry underlying male courtship rituals, fruM appears to initiate cell-autonomous male-typical differentiation of the neurons in which it is expressed. This leads to overt neuroanatomic dimorphism in cell number or projections of several specific classes of neurons (Figure 58–6B). Many neurons that express fruM are not distributed in dimorphic patterns. In these neurons fruM may regulate the expression of particular classes of genes whose products drive a male-specific program of physiology and function.
Are neurons that express fruM required for male courtship behavior? When synaptic transmission is genetically blocked in these neurons in adult males all components of courtship behavior are abolished. Importantly, these males continue to exhibit normal movement, flight, and other behaviors in response to visual and olfactory stimuli. These findings demonstrate that fruM appears to be expressed in a neural circuit that is essential for and dedicated to male fly courtship.
Control of male courtship in the fruit fly.
A. Male Drosophila melanogaster (labeled with asterisk) engage in a stereotyped sequence of behavioral routines that culminate in attempted copulation. The male fly orients toward the female and then taps her with his forelegs. This is followed by wing extension in the male and a species-specific pattern of wing vibrations that is commonly referred to as the fly courtship song. If the female fly is sexually receptive, she slows down and permits the male to lick her genitalia. The female then opens her vaginal plates in order to allow the male to initiate copulation. All steps in the male mating ritual require the expression of a sex-specific splice variant of the fruitless (fru) gene. (Modified, with permission, from Greenspan and Ferveur 2000.)
B. The fru gene encodes a male-specific splice variant that is necessary and sufficient to drive most steps in the male fly courtship ritual. Fru expression is visualized using a fluorescent reporter protein (green) in transgenic flies. Fru-expressing neuronal clusters are present in comparable numbers in the central nervous system of both male and female flies. However, there are sex differences in Fru expression as well. A cluster of Fru-expressing neurons is present in the male optic lobes (in the area within the white ellipses) but absent in the corresponding regions in the female brain. The two male antennal lobe regions (areas within yellow ellipses) contain about 30 neurons each, whereas each female region has only 4–5 neurons. (Modified, with permission, from Kimura et al. 2005.)
A Sexually Dimorphic Neural Circuit Controls Erectile Function
The lumbar spinal cord of many mammals, including humans, contains a sexually dimorphic motor center, the spinal nucleus of the bulbocavernosus (SNB). Motor neurons in the SNB innervate the bulbocavernosus muscle, which plays an important part in penile reflexes in males and vaginal movements in females.
In adult rats the male SNB contains many more motor neurons than the female SNB. In addition, male SNB motor neurons are larger in size and have larger dendritic arbors, with a corresponding increase in the number of synapses they receive. Like the SNB motor neurons, the bulbocavernosus muscle is larger in males than females; it is completely absent in the females of some mammalian species. SNB motor neurons also innervate the levator ani muscle, which is involved in copulatory behavior and is also larger in males than females.
How do these differences arise? Initially the circuit is not sexually dimorphic. At birth male and female rats have similar numbers of neurons in the SNB and similar numbers of fibers in the bulbocavernosus and levator ani muscles. In females, however, many motor neurons in the SNB and many fibers in the bulbocavernosus and levator ani muscles die in early postnatal life. Thus this sexual dimorphism arises not by male-specific generation of cells but rather by female-specific cell death.
Perinatal injections of testosterone or DHT can rescue a significant number of the dying neurons and muscle fibers in the female rat. Conversely, treatment of male pups with an androgen receptor antagonist increases the number of dying neurons and muscle fibers. So at a deeper level we see that the dimorphism results from male-specific preservation of motor neurons and muscle fibers that would die in the absence of hormone.
Where does testosterone act to establish this structural dimorphism? Is it primarily a survival factor for the motor neurons, with muscle fibers dying secondarily because they lose their innervation? Or does testosterone act on muscles, which then provide a trophic factor to support the survival of SNB motor neurons? This issue has been examined in rats carrying a mutation of the androgen receptor (tfm allele) that reduces binding of ligand to 10% of normal. The receptor resides on the X chromosome, so all males that carry a mutant gene on their one and only X chromosome are feminized and sterile. For female heterozygotes, the situation is more complicated. As described earlier, one of the X chromosomes is randomly inactivated in each XX female.
Female heterozygotes are therefore mosaics: some cells express a functional androgen receptor allele, others the mutated allele. Each muscle fiber has many nuclei, so most bulbocavernosus muscle fibers in the heterozygous female express functional androgen receptors. Motor neurons have a single nucleus, however, so each neuron is either normal or receptor-deficient. If androgen receptors were required in the neuron, one would expect only receptor-expressing SNB motor neurons to survive, whereas if receptors were required only in muscles, one would expect surviving motor neurons to be a mixture of wild type and mutant.
In fact, the latter situation occurs, indicating that survival of SNB motor neurons does not depend on a neuron-autonomous function of the androgen receptor. Rather, these neurons receive a trophic cue from the androgen-dependent bulbocavernosus and the levator ani muscles (Figure 58–7A). These cues may include the ciliary neurotrophic factor (CNTF) or a related molecule, because mutant male mice lacking a CNTF receptor exhibit a decreased number of SNB motor neurons, typical of females.
Sexual dimorphism in the spinal nucleus of the bulbocavernosus muscle.
A. The spinal nucleus of the bulbocavernosus (SNB) is found in the male lumbar spinal cord but is greatly reduced in the female. The motor neurons of the nucleus are present in both sexes at birth but the lack of circulating testosterone in females leads to death of the SNB neurons and their target muscles. It is thought that testosterone in the male circulation promotes the survival of the target muscles, which express the androgen receptor. In response to testosterone the muscles provide trophic support to the innervating SNB neurons. This muscle-derived survival factor is likely to be ciliary neurotrophic factor or a related member of the cytokine family. Thus testosterone acts on muscle cells to control the sexual differentiation of SNB neurons. (Reproduced, with permission, from Morris, Jordan, and Breedlove 2004.)
B. Dendritic branching of SNB neurons is regulated by circulating testosterone in adult male rats. In males the dendrites arborize extensively within the spinal cord (upper photo). The fact that the arbors are pruned in adult castrated male rats (lower photo) is evidence that this dendritic branching depends on androgens. The spinal cord is shown in transverse section and the SNB neurons and their dendrites are labeled by a retrograde tracer injected into target muscles. (Reproduced, with permission, from Cooke and Woolley 2005.)
Male and female SNB motor neurons also differ in size. Androgens determine the differences in number and size of these neurons in different ways. Studies of tfm mutants showed that androgens exert an organizational effect during early postnatal life through a direct effect on muscle. Low levels of androgens during this critical period lead to an irreversible reduction in the number of SNB motor neurons. Later, androgens act directly on SNB motor neurons to increase the extent of their dendritic arbors. A loss of circulating testosterone, such as that occurring after castration, leads to a dramatic pruning of dendritic arbors; injection of supplemental testosterone to a castrated male rat can restore this dendritic branching pattern (Figure 58–7B). This effect persists in adulthood and is reversible, so it can be viewed as an activational influence. Thus androgens can exert diverse effects, even on a single neuronal type.
A Sexually Dimorphic Neural Circuit Controls Song Production in Birds
Several species of songbirds learn species-specific vocalizations that are used for courtship rituals and territorial marking (see Chapter 60). A set of interconnected brain nuclei controls the learning and production of birdsong (Figure 58–8A). In some songbird species both sexes sing and the structure of the song circuit is similar in males and females. In other species, such as zebra finches and canaries, males alone sing. In these species several song-related nuclei are significantly larger in the male than in the female.
Sexual dimorphism in the avian song circuit.
A. Songbirds have a dedicated neural circuit for song production and learning, with distinct components contributing to learning or production. Many of these components are sexually dimorphic in songbirds in which only one sex sings. For example, in zebra finches the male sings, and the male high vocal center (HVC), robust nucleus of the archistriatum (RA), lateral magnocellular nucleus of the anterior neostriatum (LMAN), and area X are larger in volume and contain more neurons than the comparable regions in the female. (DLM, medial nucleus of the dorsolateral thalamus; nXIIts, hypoglossal nucleus.) (Reproduced, with permission, from Brainard and Doupe 2002.)
B. In the male the axons of HVC neurons terminate on neurons in the RA nucleus, whereas in females the axons terminate in a zone surrounding the nucleus. The sexual dimorphism in cell number and connectivity of these regions is regulated by estrogen. (Reproduced, with permission, from Morris, Jordan, and Breedlove 2004.)
C. The pattern of termination of the axons of HVC neurons in the RA nucleus varies in males and females at different ages after hatching. (Reproduced, with permission, from Konishi and Akutagawa 1985.)
The development of sexual dimorphism in song circuitry has been studied in detail in the zebra finch. In the adult male zebra finch the robust nucleus of the archistriatum (RA) contains fivefold more neurons than does the same nucleus in females. In addition, the afferent projections to RA exhibit a striking sexual dimorphism—only in males does the RA receive input from high vocal centers (HVCs) (Figure 58–8B). These sex differences in cell number and connectivity of RA are not evident until after hatching, when in females a large number of RA neurons die and in males the axons of HVC neurons enter the RA nucleus.
These sexually dimorphic anatomical features are regulated by steroid hormones. When females are supplied with estrogen (or an aromatizable androgen such as testosterone) after hatching, the number of neurons in the RA and the termination pattern in the nucleus are similar to that of the male. However, early hormone administration to young females is not sufficient to masculinize the song nuclei to a size comparable to that of adult males, nor is it sufficient to induce singing in females. To achieve these functions, female birds that receive testosterone or estradiol after hatching must also receive testosterone or dihydrotestosterone (but not estrogen) as adults. Thus steroids play both organizational and activational roles in this system.
A Sexually Dimorphic Neural Circuit in the Hypothalamus Controls Mating Behavior
In many mammalian species the preoptic region of the hypothalamus and a reciprocally connected region, the bed nucleus of the stria terminalis (BNST), play important roles in sexually dimorphic mating behaviors. In male rodents and monkeys these areas are activated during mating behavior; surgical lesions that ablate the preoptic region or the BNST result in deficits in male sexual behavior in male rodents, and also disinhibit female-type sexual receptivity. Thus this region contains neurons that activate and inhibit female sexual behavior. Surgical lesioning of the preoptic hypothalamic region activates male mating routines and inhibits female sexual receptivity in rodents.
Both the preoptic hypothalamus and the BNST are sexually dimorphic, containing more neurons in males compared to females. The sexually dimorphic nucleus of the preoptic area (SDN-POA) also contains significantly more neurons in the male. A male-specific perinatal surge of testosterone promotes survival of neurons in the SDN-POA, whereas in females these same cells gradually die off in the early postnatal period. This development is similar to that in the sexually dimorphic nuclei of the rodent spinal cord and the songbird brain, suggesting that androgen control is a common mechanism for production of sex differences in the size of neuronal populations.
Curiously, the ability of brain testosterone to promote the survival of neurons is likely to be exerted via aromatization into estrogen and subsequent activation of the estrogen receptors (see Figure 58–5). How, then, is the neonatal female brain shielded from the effects of circulating estrogen? In newborn females there is very little estrogen in the circulation, and the small amount present is sequestered by binding to α-fetoprotein, a serum protein. This explains why female mice lacking α-fetoprotein exhibit male-specific behaviors and reduced female-typical sexual receptivity. In this case, then, structural sexual dimorphism does not result from differential effects of androgens and estrogens, but rather from sex differences in the level of hormone available to the target tissue.