| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
REVIEW |
1 Departments of Basic Medical Sciences, 2 Clinical Developmental Sciences, St George’s, University of London, Cranmer Terrace, Tooting, London SW17 0RE, UK
Correspondence should be addressed to C A Wilson; Email: cwilson{at}sgul.ac.uk
 |
Abstract |
|---|
 Top Abstract Sexual differentiation of the...Sexual differentiation of the...AcknowledgementsReferences |
|---|
|
Sexual differentiation of the reproductive system |
|---|
|
TopAbstractSexual differentiation of the...Sexual differentiation of the...AcknowledgementsReferences |
|---|
This basic pattern of hormonal regulation of sexual development has been largely confirmed over the ensuing years, with some extensions and modifications that will be detailed later in this review. Advances in molecular biology have recently highlighted the importance and complexity of chromosomal and genetic regulation of sexual differentiation. Therefore, the regulation of sexual differentiation can be subdivided into chromosomal, genetic and hormonal stages, each influencing the next and resulting not only in sex differences in the reproductive system, but also in the brain. Together, the sex differences in the morphological and functional phenotype of the body and brain underlie gender identity, sexual orientation, sexual behaviour and differences in certain non-reproductive behaviours and cognitive functions. In this review, the genetic as well as a more detailed description of the hormonal control of the sexual differentiation of the reproductive system and brain will be presented. The role of certain neurotransmitters in the regulation of sexual differentiation will also be considered. This review largely focuses on experiments carried out on rodents, but wherever possible, comparison is made with data from humans.
Development of the male reproductive system
Chromosomal sex
In mammals, sexual differentiation fundamentally results from chromosomal differences, because males (typically 46XY) and females (46XX) are defined by the presence or absence of a Y chromosome (see Fleming & Vilain 2004). Genetic sex is determined at the time of fertilization by the entry of a X or Y chromosome from the sperm pronucleus into the pronucleus of the oocyte (Marshall Graves 2000). The mammalian X chromosome is large, bearing 3000–4000 genes and comparison with primitive mammals such as monotremes (egg-laying mammals), indicates that the genes on the long arm of the X chromosome (the X conserved region) are 170 million years old. The rest of the X chromosome has been added more recently and is perhaps derived from chromosomes 1 or 5 (the X added region). Evidence exists to suggest that the X and Y chromosomes were originally homologous, but the Y chromosome has been degraded over 200 million years of mammalian evolution, such that both the Y conserved and Y added regions are greatly reduced in size and gene content. The human Y chromosome now bears only 33 genes, with the rest of the chromosome comprising repetitive sequences with unknown function. Only the two tips of the Y chromosome (one bearing 9 and the other 4 genes; pseudoautosomal regions 1 and 2 respectively) are homologous with similar parts of the X chromosome. The majority of genes on the Y chromosome are expressed in the embryonic genital ridge and are concerned with development of the testis and spermatogenesis (Marshall Graves 2000).
Genetic sex
The process of sexual differentiation begins almost immediately after conception, because as early as day 2 post-fertilization, male embryos comprise more cells and have higher metabolic activity than female embryos. Sex differences in size (males>females) are present in the blastocyst (mice) and zygote (humans), as a result of the accelerating effect of the Y and the retarding effect of the X chromosomes (Mittwoch 2000). However, in the early stages of development, the foetus’ phenotype is reproductively bipotent, because the genital ridge is undifferentiated and can develop into testis or ovaries. The ridge itself develops under the influence of several genes e.g. GATA4, LHX9, L1M1 and particularly WT1 (Wilms Tumour 1) and SF1 (steroidogenic factor gene 1) both of which can be detected earlier in development than the others (Viger et al. 2005). n.b. we are following the convention of using upper case letters for human genes and lower case for rodent genes.
The defining sex-determining event is the differentiation of the genital ridge into testis with the formation of Sertoli cells, which only occurs in the presence of a Y chromosome. Testicular differentiation is tightly correlated with the high level of cell proliferation that occurs in the male embryo (Mittwoch 2000). In the early 1950s, the existence of a testis determining factor (TDF) on the Y chromosome, which initiates the masculinization of the genital ridge was predicted (see Jost et al. 1973). Two candidate genes were SRY and ZFY, with the former being confirmed as TDF following mutation analysis in humans and due to its wide distribution across species (Maxson 1997, Marshall Graves 2000). SRY was first characterized and cloned by Sinclair et al.(1990). SRY can be detected in the genital ridge soon after its formation and SRY expression increases under the positive control of WTI, SFI and GATA4 amongst many others genes (Fleming & Vilain 2004). Once sexual differentiation of the genital ridge has occurred, SRY levels fall (Koopman 2001, Park & Jameson 2005). In mice, the Sry gene was initially thought to be selectively expressed in the genital ridge, but more recently its expression has also been demonstrated in the brain (Mayer et al. 2000). SRY is expressed in many human tissues (Koopman 2001).
The SRY gene expresses a transcription factor containing a high mobility group (HMG) domain, which allows it to bind to and induce bending of its target DNA, presumably exposing new binding sites to other nuclear proteins. This appears to be essential for normal functioning of the SRY gene, because sexual abnormalities are associated with HMG mutations (Fleming & Vilain 2004). Sex reversal can occur if the SRY gene is transposed onto the X chromosome or if it is absent from the Y chromosome; these abnormalities can occur naturally (Vilain & McCabe 1998) and more recently have been induced in mice experimentally, resulting in XY females and XX males (Maxson 1997, Arnold et al. 2004). Although as yet unproven, it is thought that SRY together with SF1 as a synergistic factor, activates a cascade of other genes, which are required for development of the testis (Box 1
). A particularly important gene in this regard is SOX9 (SRY related homeobox protein 9), which expresses a protein similar to SRY that incorporates HMG and is found in developing Sertoli cells. In the absence of SRY, SOX9 can still induce full male differentiation and its presence is essential for normal development (Koopman et al. 2001). Another HMG protein, SOX8, acts similarly to SOX9, but is not essential for normal development (Koopman 2005). An example of one of the genes activated by SRY is the intragonadal regulator fibroblast growth factor (FGF) 9, part of a large family of signalling molecules. The effects of fgf9 deletion in mice indicates its importance in the proliferation and differentiation of Sertoli cells, which in turn are important in initiating spermatic cord formation and increasing Leydig cell and perituberal myoid cell numbers (Chan & Rennert 2002). In contrast to suggesting that SRY stimulates genes necessary for development of the testis, an alternative ‘disinhibition’ theory has been proposed that states that SRY encodes a repressor agent, which inhibits an inhibitory ‘Z’ factor (see Park & Jameson 2005 and Box 2).
| Box 1 Functions of genes implicated in sexual differentiation of the reproductive system Capital letters are used for human genes and lower case for other species. The functions of the SF1 (sf1), WT1 (wt1), SRY (Sry), SOX (sox9) DAX1 (dax1), AMH (amh) and WNT4 (wnt4) genes are described in the text.
References Koopman (2001), Chan & Rennert (2002), Fleming & Vilain (2004), Park & Jameson (2005).
|
| Box 2 SRY as a suppressor of a masculinizing inhibitory factor SRY is a homologue of the X chromosome gene, SOX3 and sequence comparisons between species suggest that SRY is a degraded relic of SOX3. SRY appears to have opposite, or, at least, different actions to SOX3 and some authors suggest that rather than ‘switching on’ testis differentiating genes, SRY represses a testis inhibitor, designated the ‘Z’ factor. This inhibitor suppresses pro-testis events acting on both XX and XY backgrounds upstream to SRY. Loss or mutation of the ‘Z’ factor could result in a XX male and increased ‘Z’ factor activity could cause a sex reversal (XY female) if it was high enough to overcome SRY suppression. Wnt 4 and DAX1 are possible candidates for the ‘Z’ factor since both prevent testis formation. References Vilain & McCabe (1998), Marshall Graves (2000), Bowles & Koopman (2001), Koopman (2001).
|
DAX1 (dosage-sensitive sex-reversal, adrenal hyperplasia congenital X chromosome 1) is situated on the X chromosome and is upregulated by WT1 and SF1. DAX1 regulates testis development in a dose- and time-sensitive way (Vilain & McCabe 1998, Fleming & Vilain 2004). It is essential for early testicular development, because in its absence, Sertoli cells do not develop and spermatogenesis is impaired and DAX1 is also necessary for Leydig cell proliferation and differentiation (see Sharpe 2006). However, its over-expression, e.g. when the gene is duplicated, interferes with testis development, because it encodes for a transcriptional repressor that suppresses SRY, SFI and SOX9 in a dose-dependent manner (Vilain & McCabe 1998, Park & Jameson 2005). DAX1 is present in both sexes until mid-pregnancy (embryonic day (E) 12.5 in mice), after which it is only expressed in the female and may be important for ovarian development or rather protection against testis formation in the female (Vilain & McCabe 1998). The possible sites of action and interactions of some of the genes regulating development of the male reproductive system are illustrated in Fig. 1.
|
The testis synthesize testosterone in Leydig cells from E15 in the rat (El-Gehani et al. 1998a) and 12–17 weeks of pregnancy in humans (Diez d’Aux & Pearson Murphy 1974). Low levels of testosterone are synthesized by Leydig cells in the foetal rat testis on E15 and its intratesticular concentration rises slowly over E16–17. Testicular testosterone levels then rise sharply on E18, to peak at E19 and then decline over E20–21 (Bentvelsen et al. 1995). Circulating levels of testosterone are first detectable on E16 or 17, peak at E18 and fall on E20 (Weisz & Ward 1980, Ward et al. 2003), followed by a second surge on the day of birth (Rhoda et al. 1984). Androgen receptor immunoreactivity (AR-ir) first appears in the rat urogenital tract on E14, with equal densities in males and females (Bentvelsen et al. 1995). Testicular androgens increase AR density in males over the ensuing prenatal period, resulting in a greater expression in males compared with females (Bentvelsen et al. 1995).
Leydig cells in the foetal testis are morphologically and functionally different from those in the adult (Huhtaniemi & Pelliniemi 1992). In foetal Leydig cells, the initiation of testosterone synthesis is independent of LH secretion, because LH only becomes detectable in the pituitary on E16 and in the plasma on E17. LH does not become functionally steroidogenic until E20, which is well after the appearance of testosterone on E15 and its surge on E18 (El-Gehani et al. 1998a). The early independence of testosterone secretion from gonadotrophin activity is supported by the fact that genetic mutant or knock-out male mice deficient in gonadotrophin-releasing hormone, luteinizing hormone (LH) or LH receptors, all exhibit normal prenatal masculinization (El-Gehani et al. 2000, Zhang et al. 2001). Possible alternative initiators of foetal testosterone secretion are pituitary adenylate cyclase-stimulating polypeptide (El-Gehani et al. 2000), vasoactive intestinal peptide (El-Gehani et al. 1998b) and the natriuretic peptide hormones (El-Gehani et al. 2001). These peptides and their receptors are all present around E15 in the rat and are all steroidogenic.
The synthesis of testosterone from cholesterol involves trafficking between mitochondria and smooth endoplasmic reticulum, mainly via P450 haem-containing enzymes. The genes coding for these enzymes are abbreviated to CYP and are upregulated by SF1 and GATA4. The principal enzymes involved in the production of testosterone are cytochrome P450 side-chain cleavage enzyme, the 3ß-hydroxysteroid dehydrogenases, types 1–6, the 17ß hydroxylase 17–20 lyase enzyme and the 17 hydroxysteroid reductases, types 1–3 (Stocco & McPhaul 2006); further details are given in Box 3.
| Box 3 Testosterone synthesis There are two possible pathways for testosterone synthesis:
The
The The enzymes required for these conversions are:
|
Testosterone is essential for the morphological differentiation of the male reproductive tract. It acts in an exocrine manner, diffusing down a gradient from the testis via efferent ductules to the lumen of the Wolfian tract. Thus, masculinization occurs sequentially; first the efferent ductules, then the Wolfian tract followed by the upper part of the epididymis and finally the remaining epididymis, vas deferens and seminal vesicles (Bentvelsen et al. 1995). While testicular development is initially independent of steroids, the later proliferation of Sertoli cells requires testosterone (Sharpe 2006). Testosterone is also converted to the tenfold more potent androgen, dihydrotestosterone (DHT; by 5 
-reductase type II), which, in turn, is essential for development of the male urethra, prostate gland, penis and scrotum, together with the growth of tissue between the anus and the genital orifice. The greater anogenital distance in the male is widely used as a morphological marker for differentiating the sex of neonatal rodents. SF1 induces expression of insulin-like factor 3 (INSL3) in Leydig cells, which is essential for the development of the gubernaculum and the early (intra-abdominal) stage of testicular descent. The later (inguino-scrotal) stage of testicular descent is dependent upon testosterone (Viger et al. 2005, Sharpe 2006).
Development of the female reproductive system
Genetic sex
The development of the female reproductive tract has received much less attention than the male tract, probably because of the widespread perception that it occurs ‘by default’. The equivalents of the SRY and SOX9 transcription factors have not been demonstrated in the female. However, repressor proteins preventing male differentiation are expressed in the ovary, e.g. the DAX1 gene product, which suppresses SF1 and SOX 9 activity. Another example is WNT 4 (a member of the Wnt/Winglass family), which in common with DAX1 is expressed solely in the ovary after mid-pregnancy, where it maintains oocyte numbers. In earlier pregnancy, WNT 4 suppresses Leydig cell differentiation and the synthesis of testosterone, perhaps by repressing SF1 function or inhibiting migration of steroidogenic precursor cells into the developing ovary (Fleming & Vilain 2004). WNT 4 is a candidate for the ‘Z’ factor (see Box 2), because female Wnt 4-null mice develop virilised gonads and Wolfian derivatives (see Park & Jameson 2005). Normally, Wnt 4 is down-regulated in males, thus removing any potential inhibition of SF1 function (Park & Jameson 2005).
In the female, FOXL 2 and FIG X are transcription factors that control the development of granulosa cells and oocytes within the ovary (Loffler & Koopman 2002, Fleming & Vilain 2004). In the absence of androgens and AMH, the Wolfian Tract regresses and the Mullerian tract develops into the uterus, cervix and upper vagina. DAX1 plays a role in the development of the female reproductive tract in that it inhibits AMH expression (presumably by suppressing SRY, WT1 and SF1) and WNT 4 is part of a cascade of genes controlling the development of the Mullerian tract into the oviducts, uterus, cervix and vagina (Rey et al. 2003). Animal studies have demonstrated that Wnt 4 is regulated by Wnt 7a and in turn, Wnt 4 stimulates a family of Hox a and Hox d genes. The whole pathway is down-regulated in the absence of Wnt 7a, leading to abnormalities in the development of the rostro–caudal axis of the female reproductive tract, i.e. uterine morphology develops in the oviduct, vaginal morphology develops in the uterus, the cervix and vagina fail to fuse in the midline and remnants of the Wolfian tract remain. Oestrogen down-regulates Wnt 7a and also Hox a10 at the other end of the cascade and it is likely that the abnormalities in the reproductive tract of female offspring of women treated prenatally with the synthetic oestrogen, diethylstilboestrol, are due to the oestrogenic effects of this potent drug (Kitajewski & Sassoon 2000).
Hormonal sex
In the early stages of ovarian development, the germ cell precursors (gonocytes) rapidly proliferate by mitotic division over E14–15 and this proliferation is closely followed by meiotic division, which, in contrast to the male, occurs during foetal life and is considered to be one of the first stages of ovarian differentiation (Hilsher et al. 1974). The earliest meiotic divisions occur in the germ cells close to the mesonephric connections in the central area of the ovary and it has been suggested that the mesonephros secretes a meiosis-inducing substance (Byskov 1986). This substance may be the same as the meiosis-activating sterols produced in adult gonads that are intermediaries in the cholesterol biosynthesis pathway (Byskov et al. 1998, Rozman 2000). The foetal production of the meiosis-inducing substance is not sexually differentiated, but in the male, germ cells are protected from its action by their enclosure within the testicular cords, whereas in the female, germ cells are exposed to its influence (Byskov 1986).
Activin may be involved in the regulation of germ cell division and mesonephron activity in both males and females. The activin ßB subunit and activin receptor type IIB mRNA are expressed in the gonads and mesonephroi of both sexes over E14–15 (Roberts et al. 1991, Kaipia et al. 1994). Therefore, activin and its receptors are probably present at the relevant time for female germ cell division. Experiments employing exogenous recombinant activin-A have revealed that activin stimulates DNA synthesis in the gonads and mesonephroi in a sexually differentiated manner. In males, activin inhibits DNA synthesis (see section: Development of the male reproductive system; Hormonal sex), whilst in females, it stimulates DNA synthesis, particularly over E14–15 (the time of rapid germ cell division), but activin becomes ineffective on E18 when meiosis ceases. It is therefore possible that activin stimulates division of female germ cells and also enhances the production of the meiosis-inducing substance from the mesonephron (Kaipia et al. 1994). Although follistatin has effects on adult gonads and is present in the foetal and neonatal ovary, its role in development has not been clarified (Kaipia et al. 1994).
Effects of environmental endocrine agents on sexual differentiation of the reproductive system
Compounds that can mimic or antagonise the effects of oestrogen and/or androgen are collectively referred to as endocrine disruptors (EDs). The oestrogenic agents comprise xenoestrogens, which are non-steroidal synthetic compounds used in industry and agriculture, and phytoestrogens, which are non-steroidal compounds that occur naturally in plants and fungi (see Safe 2005 and Box 4 for selected EDs and their actions in rodents). EDs can be ingested in the diet or inhaled from the atmosphere, and there are currently about 60 chemicals that have been identified and characterized as EDs (Brevini et al. 2005, Safe 2005).
Box 4 Actions of some endocrine disruptors on rodents
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
The disastrous effects of treating pregnant women with the potent non-steroidal synthetic oestrogenic compound, diethylstilboestrol, on reproductive tract development in both male and female offspring leading to testicular or vaginal cancer in adulthood, first highlighted the effects of exogenous endocrine agents on foetal development (see Bay et al. 2006). This, together with the increased prevalence of male reproductive disorders over recent years, has led to the hypothesis that environmental EDs can cause abnormal sexual differentiation of the reproductive tract and brain, and malfunction of the adult reproductive system (Colborn et al. 1993).
There are four principal disorders of the reproductive system in males: cryptorchidism (failure of testicular descent), hypospadias (ventral opening of the urethra on the penis), low sperm count (there has been a 40% decrease in sperm count in the last 50 years according to meta-analysis of 61 Scandinavian and Northern European studies (Carlsen et al. 1992)) and testicular cancer in adulthood. Together, these four disorders are referred to as testicular dysgenesis syndrome (TDS) or developmental oestrogenic male phenotype (Safe 2005, Bay et al. 2006, Sharpe 2006). Although the association between EDs and TDS has been taken seriously by regulatory bodies, whether the association is causal or not is still a matter of debate, because the expression of some aspects of TDS does not vary with changes in exposure to EDs and there are marked demographic differences in the prevalence of TDS that cannot be explained by differences in exposure to EDs. For example, there is four times more cryptorchidism and testicular cancer in Denmark than in Finland, despite similar exposure to an ED (in this case organochloride pesticides; Safe 2005). In a critical survey of studies of the reproductive health of offspring of parents exposed to EDs, Vidaeff & Sever (2005) concluded that there was not a strong correlation, but sufficient to warrant further investigation.
The evidence for an association between environmental EDs and impaired sexual differentiation and development is much stronger for wildlife than it is for humans. A wide variety of EDs have been implicated in the feminization of male birds, alligators and fish (see Hoyer 2001). Moreover, laboratory studies have replicated the effects of EDs on the development of the male rodent reproductive system that have been observed in humans. For example, administration of phthalates (which reduce testosterone synthesis) to a pregnant dam produces TDS in male offspring, similar to that observed in humans (Sharpe 2006). This has led to the suggestion that TDS is caused by low-testosterone activity over the critical period for development of the testis and differentiation of the Wolfian tract (Sharpe 2006). Once the testis have been formed from the gonadal ridge, they produce testosterone, which normally has a positive feedback effect on further testicular development. Failure of this feedback results in disordered testicular structure, abnormal sperm production and reduced testosterone synthesis. Exogenous oestrogenic agents (e.g. diethyl-stilboestrol or alkyl phenols) could induce these abnormalities by: (a) impairing Leydig cell development, (b) exerting a negative feedback effect on the follicle-stimulating hormone (FSH) release necessary for normal Sertoli cell development, (c) inactivation of steroidogenic enzymes or (d) inhibiting expression of androgen receptors (AR). Anti-androgen agents would be expected to have similar effects by inhibiting testosterone synthesis (e.g. by phthalates) or activity (e.g. by polychlorinated biphenyls; Vidaeff & Sever 2005, Sharpe 2006).
Most EDs exert weak endocrine activity, but their relatively potent biological effects can be explained by the fact that they have a low affinity for serum steroid-binding proteins and therefore, their biologically active concentration is proportionately higher than that of similar amounts of endogenous hormones that have a higher affinity for serum steroid binding proteins (Vidaeff & Sever 2005). Many EDs are also lipophilic and so bioaccumulate in fat deposits, giving rise to high body concentrations that can be released over a long period of time. EDs can enhance the activity of endogenous steroids by preventing their catabolism by enzymes such as sulphotransferase. Many EDs exhibit U-shaped dose–response curves, so that they are most potent at low concentrations. It is also possible that weak EDs may have relatively greater effects on the undeveloped foetus compared with the adult. Finally, some EDs have a wide profile of activity, e.g. bisphenyl A is an oestrogen agonist and a potent androgen antagonist, and alkyl phenols are weak oestrogen agonists and potent anti-progestagens. Both types of compound can also act as agonists or antagonists depending on their absolute concentration or their concentration relative to endogenous steroids. All of these factors could play a part in the way, in which EDs cause disruption of normal sexual differentiation (Brevini et al. 2005, Vidaeff & Sever 2005).
|
Sexual differentiation of the brain |
|---|
|
TopAbstractSexual differentiation of the...Sexual differentiation of the...AcknowledgementsReferences |
|---|
Genetic control of sexual differentiation of the brain
Genetically controlled brain functions and sexual dimorphism (hormone independent)
One of the earliest observations of genetic-dependent and hormone independent sex differences in brain tissue was reported by Reisert and colleagues (see Reisert & Pilgrim, 1991) working on diencephalon and mesencephalon cell cultures taken from E14 (endogenous gonadal testosterone is not secreted until E16 or 17) male and female rat foetuses. These authors observed that mesencephalic cultures from female rats contained more tyrosine hydroxylase (TH)-immunoreactive (ir) neurones (TH is a marker of the presence of dopamine; DA) than cultures from males. However, in males mesencephalic DA neurones were more mature than in females as indicated by increased [H3] DA uptake. In the diencephalon cultures, although the number of TH-ir neurones was not sexually differentiated, neurones from females were more mature than those from males and contained higher concentrations of DA and their TH exhibited a greater activity as assessed by the accumulation of dihydroxyphenylalanine in the presence of an aromatic amino-acid decarboxylase inhibitor. These sex differences were not affected by the manipulation of sex steroid activity (by administration of oestradiol or androgen antagonists), either in vivo before foetuses were removed from the uterus, or to cell cultures in vitro.
Deletion of the Sry gene from male mice (XY–) results in a female phenotype. When a Sry transgene is inserted into an autosome of these mice, XY– Sry mice are created, which possess testis and are fertile males. Mating XY– Sry with normal XX females produces four types of progeny: XX and XY mice without the Sry gene that bear ovaries (females) and XX and XY mice with the Sry gene that possess testis (males; Arnold et al. 2004). Comparison of these mice allows assessment of sex chromosome effects independent of gonadal sex on sexually differentiated structures and functions. Cell cultures from E14 mesencephalon revealed that XY mice with or without Sry contained significantly more TH-ir neurones than XX males, XX-Sry males and normal (XX) females (Carruth et al. 2002). These findings conflict with those of Reisert & Pilgrim (1991), who reported a greater number of TH-ir neurones in XX compared with XY mesencephalic cultures at E14 and suggest that cells from genetically manipulated animals may not necessarily be similar to those from natural XX and XY animals. Studies of adult mouse substantia nigra have revealed that there are more TH-ir cells (i.e. DA neurones) in males (XY) than females (XX) and that the expression of TH-ir is regulated by Sry and is independent of gonadal hormones. This regulatory effect is indicated by the fact that Sry mRNA is present in 10% of TH-ir cells in the substantia nigra and administration of Sry antisense oligodeoxynucleotides into the substantia nigra reduces TH expression (Dewing et al. 2006). These results support those of Carruth et al.(2002) on mesencephalic cell cultures, but are in conflict with those of Reisert & Pilgrim (1991). In the diencephalon, the density of DA neurones in the anteroventricular paraventricular nucleus (AVPV) of the hypothalamus is greater in female than in male mice (Simerly et al. 1985b). This sexual differentiation may have been masked in the studies of Reisert & Pilgrim (1991), who investigated the hypothalamus as a whole and did not find any sex difference in the density of DA neurones. The sex difference in the density of DA neurones in the AVPV is dependent on neonatal steroids and not the presence of Sry (Simerly et al. 1997, De Vries et al. 2002). Sexual differentiation of male sexual behaviour, social exploration, the size of the spinal nucleus of the bulbocavernosus and the density of hypothalamic progesterone receptors (PRs) have also been studied in XX and XY mice, with and without the Sry gene. The sex differences in all of these parameters were found to be dependent upon gonadal and not chromosomal sex (De Vries et al. 2002, Wagner et al. 2004). However, vasopressin-ir fibre density in the lateral septum (which is normally higher in males) was greater in XY-Sry males compared with XX-Sry males and similarly vasopressin-ir in XY– females was greater than in XX females (De Vries et al. 2002), indicating that this parameter is dependent on chromosomal and independent of gonadal sex. Moreover, comparing the ‘sexes’ in C57/B4/6JE i-YPOS mice that produce sex-reversed XY females, revealed that the XY females exhibit a more male-like spatial perception ability compared with XX females, whilst their performance in other tests of cognitive ability did not differ (Stavnezer et al. 2000).
To date, there has been little investigation of the interaction between genetic and hormonal influences of sexual differentiation, although there is evidence on genetically controlled differences in response to steroid action. In a study of nine different strains of rats, variations in brain structure (i.e. in the sexually dimorphic nucleus of the preoptic area; SDN-POA), aromatase activity, plasma gonadotrophin concentration and male sexual behaviour have been observed, particularly in Wistar (WF/NGu) and Noble (NB/Gr) rats, but there was no significant difference in plasma oestradiol or testosterone levels (Lephart et al. 2001). The authors suggested that these phenotypic variations may be due to differences in the response to steroids, which are controlled by quantitative trait loci, i.e. differences in expression of specific genes (Lephart et al. 2001).
In summary, sex differences in the density of TH-ir cells in the mesencephalon (Reisert & Pilgrim 1991, Carruth et al. 2002, Dewing et al. 2006) and vasopressin-ir fibres in the lateral septum (De Vries et al. 2002), spatial perception ability (Stavnezer et al. 2000) and steroid sensitivity (Lephart et al. 2001) are controlled genetically, independent of gonadal hormones. It appears likely that further such examples of genetic control remain to be discovered.
Identification of the genes controlling sexual differentiation of the brain
SRY and ZFY genes are expressed in the human male hypothalamus and cortex (Mayer et al. 1998) and Sry, Sf1 and Dax 1 genes are expressed in the male mouse brain (Guo et al. 1995, Ikeda et al. 1996, Mayer et al. 2000). Thus, these genes are in a position to directly influence brain function. Sry has been shown to activate the P450 aromatase gene, which controls expression of the aromatase enzyme that causes conversion of testosterone to oestradiol within the brain (Loffler & Koopman 2002). Sry also regulates expression of TH-ir in the midbrain (Dewing et al. 2006). The Sf1 gene is required for neuronal migration into the ventromedial nucleus (VMN) of the hypothalamus. Furthermore, the distribution of cells expressing oestrogen receptors (ERs), galanin, neuropeptide Y (NPY) and 
-aminobutyric acid (GABA) is altered in mice with a Sf1 deletion (Tobet 2002).
In an attempt to identify genes involved in sexual differentiation of the brain, microarrays and RT-PCR have been applied to brain tissue obtained from E10.5 mice, well before secretion of gonadal steroids begins on E16/17 (Dewing et al. 2003). Fifty candidate genes were found to exhibit male: female differences in expression. The expression of 36 of these genes was greater in females with that of one (the inactive X specific transcript) being up to 18.5 fold higher than in males. The female: male expression ratio of the other 35 genes ranged from 1.2 to 2.4. Eighteen genes were more highly expressed in males than in females, including two located on the Y chromosome, with male: female expression ratios of 10 (DEAD box polypeptide) and 8.8 (eukaryotic translation initiation factor 2; Y) respectively (Dewing et al. 2003). Whilst the functions of these sexually differentiated genes are as yet unknown, the fact that they exist lays a foundation for further investigation into the genetic control of brain sexual differentiation and thence CNS function.
Hormonal control of sexual differentiation of the brain
Steroid control of brain-regulated functions and behaviour
Masculinization of the brain by steroids.
Early experiments on rodents demonstrated that prenatal/neonatal exposure of females to testosterone masculinizes the pattern of gonadotrophin release and sexual behaviour in adulthood. Conversely, neonatal castration induces feminization of the same parameters in males. These findings indicate that not only is the peripheral reproductive system affected by foetal testosterone, but the hypothalamic–pituitary axis is also affected (Pfeiffer 1936, Phoenix et al. 1959, Barraclough 1966).
i) Steroid action on behaviour
Box 5 lists some sexual differentiated behaviours in rodents.
Box 5 Sexually differentiated behaviours in rats
| ||||||||||||||||||||||||||||||||||||||||
ii) Steroid action on gonadotrophin release
The feedback effect of steroids on gonadotrophin release can be inhibitory (negative feedback) or stimulatory (positive feedback). Both male and female rodents respond to the negative feedback effect of steroids, but only in female (and neonatally castrated male) rodents can steroids exert a positive feedback effect. This positive feedback leads to the typical female pattern of gonadotrophin release, with cyclical surges of LH and FSH (Fink et al. 1991, see Levine 1997). Male rats cannot respond to the positive feedback effect of steroids due to the prenatal testosterone surge (E16–19) and the later postnatal testosterone/oestradiol surge (Foecking et al. 2005) and so exhibit a tonic release of gonadotrophins (Loke et al. 1992). In contrast to the situation in rodents, there is no such sexual differentiation in primates (rhesus monkeys) and both gonadectomised males and females can respond to exogenous oestradiol priming with an LH surge and castrated males can even exhibit 28-day long ovarian cycles if provided with an ovarian transplant (Karsch et al. 1973, Norman & Spies 1986). Steroids initially were considered to act predominantly at the level of the pituitary in primates (Nakai et al. 1978), but it has become clear that in both rats and monkeys of either sex, the negative and positive feedback effects of steroids are exerted at the level of both the hypothalamus and the pituitary. After suitable oestradiol priming, decreased LH release (negative feedback) results from a reduction in the amplitude of the pulsatile release of hypothalamic gonadotrophin-releasing hormone (GnRH) in rats and primates and also a reduction in GnRH synthesis in rats. The site of the inhibitory action is the medial basal hypothalamus (especially the arcuate/median eminence; ARC/ME) in both rats and primates and it may also be exerted in the rostral POA, where the GnRH neurones are located in rats (Chappel et al. 1981, see Herbison 1998). At the same time, oestradiol desensitizes the pituitary towards the action of GnRH (Aiyer & Fink 1974, Chappel et al. 1981). The positive feedback effects of steroids are also exerted at both the hypothalamus and pituitary, by increasing the GnRH pulse rate and sensitizing the pituitary’s response to GnRH (Aiyer & Fink 1974, Spies & Norman 1975, Nakai et al. 1978, Levine et al. 1985). The main sites for the positive feedback effect of oestradiol are the POA and AVPV in the rat (Goodman 1978, Simerly 2002), but such a site has not been elucidated in primates, although it may be in the medial basal hypothalamus, where in contrast to rodents, GnRH neurones are known to be located (Silverman et al. 1994, Herbison 1998). There are sex differences in the degree of response towards steroids in both the rhesus monkey and rat; the female monkey and neonatally castrated male rat are both more sensitive to the negative feedback effect of oestradiol on LH release than the intact male (Petrusz & Flerko 1965, Steiner et al. 1976) and when the steroid negative feedback effect is removed by gonadectomy, there is a more gradual rise in LH levels in females (Zanisi & Martini 1975, Hood & Schwartz 2000). However, the response of primates and rats towards the positive feedback effect of oestradiol differ. Male and androgenised female rhesus monkeys respond to the positive feedback effect of oestradiol with a larger LH surge than normal females (Steiner et al. 1976). On the other hand, oestradiol has a greater sensitising effect on the female rat pituitary response to GnRH compared with its effect on the male pituitary (Tang & Tang 1979). Thus, adult female rats in vivo and their pituitaries in vitro are two to threefold more responsive to GnRH than males and neonatally androgenised females (Nakano et al. 1976, Fink & Henderson 1977, Liaw & Barraclough 1993).
iii) Neonatal testosterone surge
Testosterone synthesized by the foetal testis between E15 and birth is essential for masculinization of the reproductive tract as described in section: Development of the male reproductive system; Hormonal sex. Masculinization of the rodent brain is predominantly due to the surge in testosterone that peaks on E18 with a second surge on the day of birth (Weisz & Ward 1980, Rhoda et al. 1984). A neonatal surge of testosterone also occurs in humans, starting 2 weeks after birth and peaking over the subsequent 2–4 months at the level normally present in adulthood. This then declines to a negligible level at 6 months of age that is maintained until puberty, when circulating testosterone concentration rises again (Forest et al. 1980). In neonatal females, testosterone levels remain approximately ten times lower than those in male neonates (Forest 1979) and even though 90% of circulating testosterone is bound, there is still an order of magnitude more free circulating testosterone in males than in females (Dixson et al. 1998).
There is no clear evidence that neonatal secretion of testosterone is involved in masculinization of the primate brain. However, it is known to be necessary for continued development of the penis and scrotum and may thus affect normal sexual activity. Suppression of neonatal testosterone in marmosets prevents normal development of the external genitalia and male sexual behaviour, but apparently not masculinization of the brain, because male sexual behaviour can be reinstated by exogenous testosterone in adulthood (Dixson et al. 1998). Thus, it appears that the mid-gestational (weeks 8–24) rather than the post-natal rise in testosterone induces masculinization of the primate brain. Although Swaab (2004) does not reject the possibility that the neonatal surge in testosterone has a masculinizing role in the human brain, he presents examples of genetic and hormonal defects that indicate mid-gestational exposure to testosterone is more important in regulation of sexual orientation and gender identity. Furthermore, Cohen-Bendahan et al.(2005) provide evidence from primate studies that suggests that there may be multiple prenatal sensitive periods when different brain regions and thus behaviours are susceptible to hormonal modulation.
iv) The critical period for sensitivity to testosterone
The responses to the prenatal/neonatal testosterone surge in rats occur over function-specific and restricted time periods. Based on experiments in which neonatal testosterone was removed from males by castration on the day of birth, or was administered on specific days over the late gestation and neonatal period to females (or neonatally castrated males), it has been shown that in rats, testosterone-induced masculinization of sexual behaviour occurs from E18 up to a period between 15 and 30 days post partum (Bloch & Mills 1995) and defeminization of sexual behaviour occurs from E18 up to day 7 post partum (Diaz et al. 1995, Rhees et al. 1997). Masculinization of the pattern of gonadotrophin release has been shown to occur from birth up to day 9 post partum and prenatal testosterone treatment (E18–22) has been reported to have no effect on the pattern of gonadotrophin release (Diaz et al. 1995, Rhees et al. 1997). However, a recent report suggests that prenatal testosterone treatment over E16–19 does masculinize the pattern of gonadotrophin release by preventing the normal ability of oestrogen to induce PRs, which in turn are essential for the induction of the pre-ovulatory LH surge. The same prenatal testosterone treatment in females also masculinizes LH pulse frequency, which is normally greater in males (Foecking et al. 2005). Traditionally, the ‘critical period’ for sexual differentiation of the brain has been described to extend from E18 (and possibly as early as E16) to approximately day 10 post partum (MacLusky & Naftolin 1981, Arnold & Gorski 1984). However, application of testosterone as late as days 15–30 post partum and removal of the testis up to day 29 post partum can influence aspects of sexual behaviour and the size of the SDN-POA (Bloch & Mills 1995, Bloch et al. 1995, Davis et al. 1995) and so for certain features the critical period may be more extensive. For others, e.g. the dark granule cell population of the accessory olfactory bulb, it may start as late as 1 or 2 weeks post partum, because administration of testosterone on day 14 post partum but not day 1 post partum can masculinize the number of these neurones (Segovia et al. 1999).
More recently, evidence has accumulated to suggest that there are two critical periods for masculinization of the brain in the rat, one over the neonatal period (possibly extending into the prepubertal period) and a second over puberty, coincident with the increase in circulating gonadal steroids. During the second period, sexually differentiated organization of the brain is completed and the behavioural potential of the adult is fully achieved. At puberty, both steroid-dependent and steroid-independent changes occur in the brain, mainly concerned with pruning synaptic connections to refine neuronal circuitry and alter neuronal responses to external stimuli. Examples include a reduction in the dendritic length of neurones in the spinal nucleus of the bulbocavernosus (which is testosterone dependent), a reduction in the size of the medial nucleus of the amygdala (which is independent of testosterone) and development of responses to female pheromones with increased dopaminergic activity in the medial POA and enhanced LH and testosterone secretion. Without these and other changes, behaviour remains in a pre-pubertal state, with low levels of male sexual behaviour, extended stress responses and unchanged levels of aggression. It has been suggested that it is important for the steroid-dependent and independent changes to occur concomitantly for full brain differentiation and that miss-timing will lead to atypical sexually differentiated behaviours and functions (Romeo 2003).
The critical period for sexual differentiation of the brain has not been defined in humans, but gender identity is thought to be established by the first year of life and structural differentiation principally occurs around 4 years of age (e.g. the size of the SDN-POA), but full sexual differentiation of the central nucleus of the bed nucleus of the stria terminalis is not manifest until adulthood (Swaab et al. 2003).
v) The relative effects of oestradiol and testosterone on sexual differentiation
In most mammals, the principal hormone masculinizing the brain is not testosterone itself, but its metabolite, oestradiol, acting on oestrogen receptors and ß (ER and ß), which control separate aspects of differentiation. ER is primarily involved in masculinization, while ERß mediates defeminization of sexual behaviours, but not masculinization (Baum 2003, Patchev et al. 2004, Kudwa et al. 2006). The oestradiol synthesized locally in the brain is derived from testosterone after aromatization by the cytochrome p450 enzyme aromatase that is present in specific brain areas and whose expression is stimulated by testosterone acting on AR (Hutchinson 1997). Some sexually differentiated behaviour in rats require testosterone as well as oestradiol for their full expression e.g. male sexual behaviour, social play, aggression and open-field behaviour (Meaney 1989, Breedlove 1997, Sato et al. 2004). While the dimorphism in most sexually differentiated brain areas is dependent entirely on oestradiol, there are a number of regions that are masculinised by testosterone or DHT acting via AR. For example, the number of neurones in the spinal nucleus of the bulbocavernosus and the connectivity between GABA nerve terminals and hypothalamic GnRH neurones are greater in males than in females (Breedlove 1997, Sullivan & Moenter 2004). In contrast, prenatal DHT treatment reduces the number of neurones in the locus coeruleus in males compared with females (Garcia-Falgueras et al. 2005). The masculinization of the pattern of gonadotrophin release initially appears to be under the control of prenatally secreted testosterone, but postnatally both testosterone and oestradiol may be involved (Foecking et al. 2005). Progesterone derived from either the maternal system or produced locally in the brain is also involved in CNS sexual differentiation and may be part of a cascade originating with testosterone and progressing to oestradiol and then progesterone to induce full masculinization (Quadros et al. 2002a).
In species, in which oestradiol is the important masculinizing agent, it is vital to protect the female foetus from the high-circulating levels of maternal oestrogen and in the rat, this is achieved by the presence of -fetoprotein secreted by the liver and yolk sac. -Fetoprotein binds to oestradiol and prevents its passage into the brain. However, it does not bind to testosterone, which freely enters the brain where it is aromatized locally to oestradiol (Naftolin et al. 1991). In a few species, e.g. the guinea-pig and rhesus monkey, testosterone (and/or DHT) has been shown to be the main masculinizing hormone of the brain (Goy & Resko 1972, Goldfoot & Van Der Werff Ten Bosch 1975). Clinical evidence from individuals with genetic disorders suggests that testosterone is also the masculinizing agent in humans (Swaab 2004). Thus, males deficient in aromatase or expressing mutant ERs, exhibit male psychosexual behaviour and gender identity, in the face of lack of oestradiol synthesis or activity. Conversely, males deficient in ARs, but with functional ERs (androgen insensitivity syndrome), not only develop an external female phenotype and feminized sexually dimorphic behaviour, but also female gender identity (Grumbach & Auchus 1999, Wilson 1999). Therefore, in humans, the absence of oestradiol activity does not prevent masculinization, but the absence of testosterone activity does. Moreover, in congenital adrenal hyperplasia, in which the foetus is exposed to high levels of adrenal testosterone due to a defective adrenal enzyme, female children display typical male activities and interests throughout life. However, in adulthood most women with congenital adrenal hyperplasia are heterosexual, although a greater percentage than in the population at large (30% compared with 10%) is homosexual, i.e. their sexual orientation is masculinized resulting in a preference for females (see Breedlove 1994, Cohen-Bendahan et al. 2005).
vi) Indirect indicators of hormonal inducement of sexual differentiation
There is general agreement that the male brain exhibits greater asymmetry than the female brain (Wisniewski 1998, Gadea et al. 2003). It is also well established that the left cerebral hemisphere typically controls speech and language, while the right hemisphere controls non-verbal (e.g. spatial perception) and emotional processing. Males typically show superiority in spatial perception compared with females (Rilea et al. 2004), while females (especially in childhood) exhibit better speech and language skills (Smith & Hines 2000, Kansaku & Kitazawa 2001). These findings support the suggestion that in males, the right cerebral hemisphere is larger than the left. In view of the fact that the left side of the body is controlled by the right cerebral hemisphere and vice versa, a dominant right hemisphere would confer left-handedness and there is evidence to suggest that more males than females are left-handed (Smith & Hines 2000, Rilea et al. 2004). However, some authors of studies of handedness conclude the opposite, i.e. that the left hemisphere is dominant in males (Witelson & Nowakowski 1991, Gadea et al. 2003, Cohen-Bendahan et al. 2004). The reason for this apparent conflict is not clear, but it is possible that handedness does not correlate with lateralization of either cortical size or function (Gadea et al. 2003).
Sexual differentiation of cerebral lateralization is thought to result from the prenatal secretion of testosterone in the male, based on clinical evidencefrom humans subjected to abnormal prenatal steroid exposure, i.e. congenital adrenal hyperplasia, prenatal exposure to diethylstilboestrol and females with opposite sex twins (Smith & Hines 2000, Cohen-Bendahan et al. 2004). Animal studies have revealed that administration of testosterone or oestradiol to developing rats induces a male pattern of cerebral lateralization, such that males develop a thicker right cerebral cortex (Diamond 1991). Witelson & Nowakowski (1991) have suggested that cerebral asymmetry is due to testosterone causing pruning of commissural fibres in the corpus callosum, leading to lateralization of cerebral functions in the male. On the other hand, the larger corpus callosum containing a greater number of axons interconnecting the cerebral hemispheres in human females potentially leads to a greater hemispheric exchange of information and therefore decreased functional asymmetry (Wisniewski 1998, Kansaku & Kitazawa 2001). There are other potential indirect measures of hormonal inducement of sexual differentiation, such as otacoustic emissions and digit length ratios (see Cohen-Bendahan et al. 2005). However, reports about these parameters are conflicting and there is no direct evidence that pre and/or perinatal testosterone is involved in these sex differences (McFadden et al. 2005, Lippa 2006).
Feminization of the brain by steroids.
Until the early 1980s, it was assumed that the brain became feminised by default in the absence of gonadal steroids. Döhler et al.(1984) questioned this assumption, when they found that administration of an anti-oestrogen compound to rats prevented normal female sexually differentiated cyclical release of LH and female sexual receptivity. They suggested that oestradiol is needed for a fully feminised brain, but in a much lower concentration than is required in the male and that there is a continuum of differentiation between the two sexes. Thus, the oestrogen requirement in male and female differentiation is quantitative rather than qualitative. A more recent finding goes further in identifying a proactive action of oestradiol on rat female brain structure. The corpus callosum (which is larger in the male rat) becomes enlarged in the female after ovariectomy in the second week of post partum and this enlargement can be prevented by the administration of oestradiol. This late effect of oestradiol contrasts with the earlier masculinizing effect on the corpus callosum observed in males before day 8 post partum (Bimonte et al. 2000).
Steroid control of sexual differentiation of brain structure
The steroid-induced structural changes in the brain are regionally specific and only occur in those brain regions expressing high densities of steroid receptors (Simerly 2002). Hormone-dependent brain sexual dimorphism was first reported as the larger size of neuronal nuclei and nucleoli in males compared with females and neonatally castratedmales (Pfaff 1966). Subsequently, Raisman & Field (1971) demonstrated a greater number of dendritic spines in the dorsal POA of the female compared with the male and Gorski et al.(1978) showed that the SDN-POA was four to six times larger in the male than female. Box 6 lists some of the many sexually differentiated nuclei in the brain.
Box 6 Sexually differentiated brain regions
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
It is now known that there is considerable variety in the nature of structural sexual dimorphisms in the brain and these include the size of specific brain regions, the extent of dendritic aborization, differences in the density and pattern of synaptic connections (e.g. spine and somatic synapses), size, number and phenotype of neurones in a particular region and astrocyte morphology (Matsumoto & Arai 1997, McCarthy et al. 2002a, Simerly 2002). In rodents, all of these differences are due to the effect of neonatal testosterone/oestradiol on programmed cell death, neurite growth, axon guidance and synaptogenesis (Fernandez-Galaz et al. 1997, Simerly 2002, see Boxes 6 and 7). Functionally, the structural sexual dimorphisms in individual brain regions give rise to sex differences in neuronal circuitry and thus differentiation of the responses to outside influences. For example, in the circuit underlying the regulation of LH and prolactin release, there are more projections from the AVPV to GnRH neurones in the POA and also to the tubero-infundibular dopaminergic system in the ARC nucleus in the female. Sensory information from the cortex can be transmitted to the hypothalamus by sexually differentiated pathways originating from vasopressin neurones in the bed nucleus of the stria terminalis and the medial nucleus of the amygdala that project to the AVPV-ARC complex controlling neuroendocrine functions and to the SDN-POA-VMN complex, concerned with reproductive behaviour (Segovia et al. 1999, Simerly 2002, De Vries & Panzica 2006, see Fig. 2).
|
Box 7 Actions of oestrogens on neuronal development, growth and survival
|
4 pathway utilised in most rodent species: Cholesterol
Pregnenolone