Clinical Neuroendocrinology

Index

Chapter 1

Gonadal Hormones and Sexual Behavior

Jean D. Wilson

Publisher Summary

This chapter discusses whether animal models of sexual behavior are applicable to the understanding of human sexual behavior. Development of reproductive capacity involves diverse activities—such as the formation of the male and female phenotypes during embryogenesis, sexual maturation, and the onset of gametogenesis at the time of puberty—and the acquisition of specific behavioral patterns, such as sexual drive and capacity for intercourse as well as patterned behavior. In all species including man, gonadal steroid hormones are involved in the conversion of the sexually indifferent embryo into the male phenotype, in sexual maturation of males and females during postnatal life, and in the development of a basic sexual drive at the time of sexual maturation. In many animal species, gonadal steroids also play a critical role in the development of the specific actions that characterize male and female reproductive behavior. It has been documented that a portion of the action of gonadal hormones in this regard is because of direct effects on the central nervous system. Steroid hormones act via a common intracellular molecular machinery involving a high-affinity receptor protein and the genetic machinery to exert their effects in such diverse tissues as external genitalia and brain.

I. Introduction

II. Normal Physiology of Gonadal Steroids

III. Phases of Sexual Life

IV. Normal Sexual Differentiation

A. Male Development

B. Female Development

C. Breast Development

D. Endocrine Control of Male Development

V. Effect of Hormones on the Sexual Behavior of Animals

VI. Control of Libido and Potentia

VII. Sexual Identification in the Human

A. Abnormal Sexual Development

B. Behavioral Studies in Patients with Abnormalities of Sexual Development

C. Apparent Reversal of Gender Identity

VIII. Management of Patients with Ambiguous Genitalia

IX. Conclusion

References

I. INTRODUCTION

Development of reproductive capacity involves diverse activities, namely, the formation of the male and female phenotypes during embryogenesis, sexual maturation and the onset of gametogenesis at the time of puberty, and the acquisition of specific behavioral patterns such as sexual drive (libido) and capacity for intercourse (potentia) as well as patterned behavior. Patterned behavior encompasses the actions that govern sexual behavior per se such as gender identity (the extent to which one perceives oneself as male or female) and gender role (the manner by which gender identity is conveyed to others) as well as actions that involve aggression, the drive for dominance, and parenting. During the past 40 years a considerable body of information has accrued to indicate that in many animal species all of these reproductive functions (phenotypic differentiation, sexual maturation, and sexual behavior) are largely mediated by steroid hormones secreted by the ovaries and testes. These steroid hormones act in the central nervous system by the same molecular mechanisms that mediate the actions of the hormones in other tissues. Some actions of steroid hormones in the brain result in permanent anatomical and functional changes, whereas other effects require continued exposure to these hormones.

In the human gonadal steroids are also responsible for phenotypic differentiation, sexual maturation and, to a large extent, development of libido and potentia. Whether gonadal hormones are also involved in the development of patterned behavior in the human male and female is not so clear-cut. It is difficult to devise appropriate experiments to establish the significance of hormonal factors in human behavior, but on the basis of studies in diverse clinical abnormalities it has been the predominant view that human reproductive behavior is determined primarily, if not exclusively, by psychologic and social factors and that biological forces (including the action of hormones) play an insignificant role in its control. According to this anthropocentric view, the phenomena that mediate behavior in animals are not adequate to explain human sexual behavior. However, a number of individuals with abnormalities of sexual development have been described who were considered female at the time of birth and were so raised but who subsequently developed partial or complete virilization at the time of expected puberty and underwent a reversal of gender role and an apparent reversal of gender identity; i.e., putative females turned into functional as well as anatomic males. Several recent reports of such patients have caused a revival of the zoocentric concept of human sexual behavior, the view that the same general biological phenomena determine animal and human behavior. These reports raise fundamental questions about the factors that regulate human sexual behavior as well as about the appropriate management of certain disorders of human sexual development.

The purpose of this review is to summarize the current dilemma as to whether the animal models of sexual behavior are applicable to the understanding of human sexual behavior. To provide the background for this problem, it is necessary first to summarize normal gonadal endocrinology and the mechanisms by which gonadal steroids act within cells. Second, since many studies of the role of hormones in human sexual behavior have utilized pathological states including abnormalities of sexual development, the role of gonadal steroids in normal sexual development is described. Third, the role of gonadal steroids on the sexual behavior of animals is reviewed.

II. NORMAL PHYSIOLOGY OF GONADAL STEROIDS

The structures of the principal gonadal steroids are shown graphically in Fig. 1. The two principal circulating androgens are testosterone and androstenedione, and the major estrogens are estrone and 17β-estradiol. In part these hormones are secreted into the circulation directly by the ovary and testis (or by the adrenal), and in part they are formed in peripheral tissues from circulating precursors (so called prohormones). Likewise, in some instances testosterone and androstenedione can serve as precursors for the formation of other steroids. For example, testosterone, the major androgen formed by the testis, serves as a precursor for the formation of the potent androgen dihydrotestosterone in androgen target tissues such as the prostate. Testosterone in turn can be formed from the weaker androgen androstenedione. Furthermore, both testosterone and androstenedione can be converted to estrogens in peripheral tissues. For example, 17β-estradiol can be formed directly from testosterone or indirectly via the sequence androstenedione → estrone → 17β-estradiol. Since both of these estrogens can bind to the estrogen receptor we will assume for the purposes of this discussion that both are active intracellular estrogens; likewise since both testosterone and dihydrotestosterone bind to the androgen receptor it is believed that they are the active androgens. The interconversions of androgens and estrogens can be demonstrated in many tissues, and the available evidence suggests that the peripheral conversion of androgen to estrogen is more important as a pathway of estrogen formation in the human than in lower animals. The quantitative aspects of this complicated metabolic pathway obviously differ under different physiological conditions, but the current view of the sources of circulating androgens and estrogens in the normal human is summarized schematically in Fig. 2. In the female about two-thirds of the estrogen formed during a menstrual cycle is derived from the ovary directly, while one-third is formed from circulating androgen. In the oophorectomized woman, therefore, estrogen production continues at about one-third the normal rate (Siiteri and MacDonald, 1973). Likewise, in the premenopausal woman, about one-half of the testosterone is secreted directly by the ovaries and adrenals, and the remainder is formed in peripheral tissues from weaker androgens; in the oophorectomized woman the adrenal component remains intact so that on the average testosterone production falls by about one-half (Kirschner and Bardin, 1972).

Fig. 1 Principal androgens and estrogens. The interconversion of these steroids is indicated by the arrows.

Fig. 2 Contribution of the gonads and the adrenals to the major circulating androgens and estrogens in the human during the reproductive years and in the castrate or postmenopausal state. The data are meant to be representative only and have been taken from Kirschner and Bardin (1972) and from Siiteri and MacDonald (1973).

In men, most estrogen (85%) is formed from peripheral conversion of circulating testosterone and androstenedione; in the castrate man adrenal production of androstenedione is unaffected, and, consequently, estrogen production is roughly one-half the normal rate (Siiteri and MacDonald, 1973). In contrast, as much as 98% of the testosterone in normal men is secreted directly by the testes so that in the castrated man testosterone production is only about 2% of normal (Siiteri and MacDonald, 1973).

To summarize, considerable estrogen production in men and women (30–50% of the total) is derived ultimately from the adrenal so that castration diminishes but does not abolish its formation. In women, a larger fraction of testosterone production (roughly one-half) is derived from the adrenal compared to men (approximately 2%). The net effect is that castration in women cuts the production of androgen and estrogen to half or less but does not have a dramatic effect on the ratio of the two steroids. In contrast, castration in men has a more deleterious effect on the production of androgens than of estrogens with the net result that castration may actually feminize men (Wilson et al., 1980).

It should also be noted that the sum of the actions of testosterone in cells includes the effects of testosterone itself, of its androgen metabolites such as dihydrotestosterone, and of its estrogenic derivatives. Since they cannot be converted to androgens, estrogens have no androgenic effects. The extent to which the physiological effects of testosterone are mediated by estrogen is one of the unsolved problems of endocrinology (Wilson, 1975); estrogenic metabolites may be major mediators of the action of androgens on the central nervous system. Indeed, estrogens formed within specific target cells may have consequences different from those of circulating estrogens.

The mechanism by which androgens exert their cellular actions within target tissues is summarized schematically in Fig. 3. Testosterone (T), the major androgen secreted by the testis and the major androgen in plasma, enters target tissues by a passive diffusion process. Inside the cell testosterone can be converted (Fig. 5) to dihydrotestosterone (D) by the 5α-reductase enzyme. Either testosterone or dihydrotestosterone is then bound to the same high-affinity androgen receptor protein (R) in the cytosol. The hormone-receptor complexes (TR and DR) move from cytosol to the nucleus. Inside the nucleus the steroid-receptor complexes interact with acceptor sites on the chromosomes and increase transcription of specific structural genes so that new messenger RNA and new proteins ultimately appear in the cytoplasm of the cell. The testosterone- and dihydrotestosterone-receptor complexes elicit distinct (separate) responses in the cell.

Fig. 3 The mechanism of androgen action. T, testosterone; D, dihydrotestosterone; R, androgen receptor protein of the cell cytosol; LH, luteinizing hormone (Wilson et al., 1981).

Fig. 5 Jost paradigm for sexual development.

The mechanism of action of 17β-estradiol and estrone is similar to that of androgen. The hormones enter target tissues by a passive diffusion process, are bound to a specific high-affinity estrogen receptor protein in the cytosol, and pass to the nucleus for interaction with acceptor sites on the chromosomes, thereby inducing transcription of specific genes. Estrogen plays a major role in the development of female secondary sex characteristics at puberty. Furthermore, just as normal female cells contain the capacity to respond to androgen, males also respond to estrogen. For example, administration of estrogen to men at any age can cause florid breast enlargement.

III. PHASES OF SEXUAL LIFE

Although they work at all stages of life by the molecular mechanisms described above, gonadal steroids have different physiological functions at different stages of life. Typical male plasma testosterone values throughout life are illustrated graphically in Fig. 4. Normally, there are three distinct periods in life in which circulating testosterone levels are elevated. During the fetal phase (the second trimester) androgens serve to promote the formation of the male phenotype. The function of androgen during the neonatal surge is not established in all species, but in rodents some permanent imprinting of androgen on the central nervous system takes place during this interval. Finally, during the adult phase of male sexual life, sexual maturation and the full capacity for reproduction are acquired. Thus, the same hormone working via the same intracellular mechanism can have different effects depending on the stage of development of the individual.

Fig. 4 Schematic diagram of the phases of male sexual function as indicated by mean plasma testosterone level and sperm production at different phases of life (Griffin and Wilson, 1979).

An analogous two-phase diagram could be written for female sexual life with ovarian estrogen formation occurring principally during fetal life and during the sexually mature phase of adult life.

IV. NORMAL SEXUAL DIFFERENTIATION

Since a major function of testosterone is to promote the development of the male phenotype during embryonic life it is necessary to describe the mechanism by which the two sexes differentiate. Male and female embryos develop in an identical fashion during the initial phases of gestation, and only thereafter do anatomical and physiological development diverge to result in the formation of the male and female phenotypes (Wilson, 1978). The fundamental mechanism of sexual differentiation was elucidated by Alfred Jost (Jost, 1972). According to the Jost formulation—now the central dogma of sexual development—sexual differentiation is a sequential, ordered, and relatively simple process (Fig. 5).

Chromosomal sex, established at the time of conception, directs the development of either ovaries or testes (gonadal sex). If testes develop, their hormonal secretions elicit the development of the male sexual anatomy, collectively known as the male phenotype. If ovaries develop or if no gonads are present, anatomical development is female in character. The mechanism by which the sex chromosomes dictate gonadal differentiation is believed to involve a specific differentiative antigen (the H-Y antigen) that is encoded in the Y chromosome and that is responsible for differentiation of the testis (Wachtel, 1980; Ohno, 1978). Regardless of the mechanisms that control histological development of the gonads, it is through their action as endocrine organs that gonads cause phenotypic and functional development of the sexes. Consequently, we will consider first the anatomical events involved in the formation of the sexual phenotypes and then the mechanisms by which gonadal hormones mediate this development.

The temporal relation between the differentiation of the ovary and testis and the development of the sexual phenotypes in the human embryo is summarized in Fig. 6. From observation of embryonic development up to about the 20-mm crown-rump (CR) length level it is evident that the gonads in the two sexes are identical and consist of three components: the primordial germ cells (which have migrated to the urogenital ridge from the embryonic hindgut), the connective tissue of the genital ridge, and a covering epithelium. Histological differentiation begins when the germ cells in the testis line up to form the so-called spermatogenic cords, a process believed to be controlled by H-Y antigen. Shortly thereafter, Leydig (interstitial) cells appear in the connective tissue of the testis; these cells synthesize testosterone. Although histological differentiation of the ovary is not apparent until the second third of gestation when the definitive ovarian follicles develop, the synthesis of 17β-estradiol by the ovaries commences at approximately the same time as testosterone synthesis in the testis.

Fig. 6 Temporal sequence of sexual development in the human embryo (Wilson et al., 1981)

Phenotypic development in both male and female embryos is largely accomplished between 30 and 90 mm CR length. The anatomical events in the development of the male and female urogenital tracts are summarized in Figs. 7 and 8. At the end of the indifferent phase of phenotypic differentiation the primordial genital tracts of both sexes consist of the gonads, two genital duct systems (wolffian and müllerian), and a common opening for the genital ducts and the urinary tract to the outside through the genital folds on the abdominal wall.

Fig. 7 Development of the internal genital tracts in the two sexes (Wilson et al., 1981).

Fig. 8 Development of the external genitalia in the two sexes (Wilson et al., 1981).

The internal genital tracts in the two sexes develop from different duct systems—the wolffian ducts persisting in the male and the müllerian ducts persisting in the female. The external genitalia in the two sexes develop from common primordia: the genital tubercle, folds, and swellings.

A. Male Development

The principal components of the male urogenital tract are shown in Fig. 7. Development of the male urogenital tract (termed virilization) commences shortly after the formation of the spermatogenic tubules of the testis. The initial event is regression of the müllerian ducts, resulting ultimately in their disappearance. Müllerian duct regression is quickly followed by virilization of the wolffian ducts. The upper portion of the wolffian duct is connected to the testis to form the epididymis, the central portion becomes the vas deferens, and the terminal portion gives rise to the ejaculatory duct and seminal vesicle. Simultaneously, the prostate gland arises from a series of buds that appear in the lining of the primitive urethra.

Development of the penis and scrotum in the male commences shortly after the onset of virilization of the wolffian ducts (Fig. 8). The genital folds elongate and fuse to form the penis and to bring the end of the urethra to the glans penis. As a consequence of the fusion the genital swellings form a bilobed scrotum that serves as the receptacle for descent of the testes.

Anatomical development of the male internal and external genitalia is accomplished largely by the end of the first trimester of gestation. In the latter part of gestation two final aspects of male development take place, descent of the testes and growth of the genitalia.

B. Female Development

In the female embryo the müllerian ducts develop into the fallopian tubes and uterus and contribute to development of the vagina (Fig. 7). In contrast, the wolffian ducts either regress or persist in remnant form. The external genitalia in the female undergo little differentiation as compared with the indifferent state (Fig. 8). The genital tubercle becomes the clitoris, the adjacent genital swellings give rise to the labia majora, and the genital folds become the labia minora.

C. Breast Development

Only one functional mammary bud develops on each side of the chest wall in the human embryo. This bud is well differentiated by the 40-mm stage, and by midgestation a nipple and secondary epithelial buds can be recognized. During the remainder of embryonic life ductular proliferation continues so that by the time of birth 15–25 branches are present. Although clear-cut sexual dimorphism in breast development occurs in some species (Kratochwil and Schwartz, 1976), such dimorphism has not been documented in the human embryo, i.e., the breasts of boys and girls are identical prior to the onset of puberty (Pfaltz, 1949).

D. Endocrine Control of Male Development

Jost established that the transformation of the indifferent urogenital tract and external genitalia into the male phenotype is determined by secretions of the fetal testis (Jost, 1972). The basis for this formulation was the demonstration that removal of the gonads from embryos of either sex prior to the onset of phenotypic differentiation results in the development of embryos with a female phenotype. Thus, the male is the induced phenotype in that testicular secretions cause formation of the male urogenital tract. In contrast, female differentiation is not dependent on the presence of an ovary and therefore does not require secretions from the embryonic gonad.

Jost also deduced that two secretions from the fetal testis are essential for male development: müllerian inhibiting substance and androgen (Jost, 1972). Müllerian inhibiting substance is an incompletely characterized protein hormone that acts in the male to cause regression of the müllerian ducts. The formation of the inhibiting substance constitutes the first endocrine function of the embryonic testis. The inhibiting substance is a glycoprotein and is formed by the spermatogenic tubules (Donahoe et al., 1977; Picard et al., 1978). Failure to produce the müllerian inhibiting substance or inability of the tissue to respond to the hormone results in a disorder (the persistent müllerian duct syndrome) in which genetic and phenotypic men have fallopian tubes and uteri together with male wolffian duct structures (Sloan and Walsh, 1976).

The second developmental hormone of the fetal testis was identified by Jost to be a steroid. The principal steroid hormone formed by the testis in fetal as well as in postnatal life is testosterone (Fig. 1); in the human embryo the fetal testis begins to synthesize testosterone shortly after the onset of differentiation of the spermatogenic tubules and concomitant with the histological differentiation of the Leydig (interstitial) cells (Siiteri and Wilson, 1974) (Fig. 6). Testosterone performs two functions in the embryo. It acts locally in the testis to promote maturation of the spermatogenic tubules and formation of sperm. Second, it is secreted into the fetal circulation where it plays an essential role in development of the male genital tract.

As the result of studies in animals and humans with single gene mutations causing resistance to androgen action and partial or complete failure of male development, it has been established that androgens act similarly in the embryo and in the postnatal state (Fig. 3) (Wilson, 1975; Griffin and Wilson, 1980). It has also been deduced that the testosterone-receptor complex is responsible for virilization of the wolffian duct, whereas the dihydrotestosterone-receptor complex induces virilization of the urogenital sinus and external genitalia during embryogenesis (Fig. 3). The reason that testosterone mediates some androgen effects while dihydrotestosterone mediates others is unclear, but the mechanism may involve a difference in the affinity of the receptor for the two androgens.

Female embryos have the same androgen receptor system and the same ability to respond to androgens as the male embryo. If a female embryo is exposed to androgens early in embryogenesis it will virilize like the male. Such androgen exposure can come from androgen ingestion by the mother, from maternal tumors or from overproduction within the fetus. Thus, the anatomical differences in the sexual phenotypes between men and women reside solely in the hormones produced by the gonads.

V. EFFECT OF HORMONES ON THE SEXUAL BEHAVIOR OF ANIMALS

The influence of gonadal hormones on behavior in animals is now well established as documented in several recent extensive reviews (Whalen, 1977; Beach, 1977; Davidson, 1972; Phoenix et al., 1967; Arnold, 1980; Resko, 1975; McEwen, 1980). Many aspects of the problem are beyond the scope of the present discussion, but several issues deserve emphasis:

1. Sexually dimorphic behavior patterns of diverse types are regulated by gonadal steroids, ranging from the songs and mating rituals of birds to copulatory patterns in mammals. For example, male and female rodents differ in the predominant type of sexual postures they assume during coitus as well as the sexual partner they pursue. These behaviors can be changed by appropriate hormonal manipulation. If testosterone is given to rats soon after birth the female is made anovulatory. Likewise, if male rats are castrated during the neonatal period, their sexual response as adults to estrogen and progesterone is like that of females.

2. Although androgens and estrogens are formed in both males and females (with both hormones possibly playing roles in the sexual physiology of both sexes) the general statement can be made that androgens (and androgen metabolites) dictate male behavior patterns and that estrogens (and to a certain extent progesterone) dictate female behavior patterns.

3. These hormones act in the central nervous system via the same molecular mechanisms that operate in peripheral tissues (McEwen, 1980; Celotti et al., 1979). Steroid hormone receptors in brain have been isolated and characterized and shown to have specific distributions within certain areas of that tissue. Pathological states in which either the hormones or the machinery of hormone action are aberrant may thus cause effects on the central nervous system as well as anatomical and functional defects in peripheral tissues. Steroid hormones may also have effects in the central nervous system in addition to those mediated by the known receptor mechanisms [e.g., effects on cell permeability (McEwen, 1980)].

4. In the rodent the neonatal surge of testosterone secretion (Fig. 5) appears to play a vital role in virilizing hypothalamic function, namely, in determining a tonic pattern of gonadotropin release as compared to the cyclic patterns in females. (This action of testosterone in the central nervous system may be mediated by testosterone metabolites such as 17β-estradiol.) A neonatal surge in testosterone secretion also occurs in the human male infant and could conceivably play a similar role.

5. Two general types of effects of gonadal steroids on behavior can be delineated. Following the terminology of Phoenix and colleagues these phenomena are termed organizational and concurrent (Phoenix et al., 1959). Organizational effects are those that require the presence of the steroid at a specific time of development, that appear to result in a permanent effect on function or behavior, and that persist (to a greater or lesser degree) even after the steroid is no longer present. Such organizational effects are thought to take place during the neonatal surge of testosterone production and may be accompanied by permanent changes in anatomical development and organization of the brain (Gorski et al., 1978). Concurrent effects require the continued presence of the steroid for the full manifestation of the effects, e.g., the mounting response of the female rodent when in estrus. Although the delineation of these phenomena is of conceptual advantage, there is considerable overlap between them in the sense that organizational effects may be silent in the absence of the proper hormonal milieu. Furthermore, some concurrent phenomena such as the typical male behavior involved in intromission and ejaculatory thrusting may persist to a variable degree in the castrated animal.

6. Virtually every behavioral effect of steroid hormones is due to complicated interaction between peripheral and central actions of the hormones (Davidson, 1972). Consider, for example, the paradigm of sexual behavior in the mammal that has been best explored, the mounting reflex of the female rat. When the female in estrus is mounted by a male, she extends the hind legs, elevates the rump, and dorsiflexes the vertebral column. This action involves not only sensory input from the rump but a well-defined neural arc that includes motor and sensory components and specific estrogen-dependent nuclei in the central nervous system. The threshold of hormone action, however, may differ in the various components of this neurogenic arc. Thus, while the central nervous system plays a vital role in the hormone-mediated control of sexual behavior in animals, individual components of behavior may be influenced to different degrees by central versus peripheral actions of the hormones. Even under defined laboratory conditions, it is difficult to devise experimental conditions that allow one to quantitate the relative contribution of each to a given action (Davidson, 1972).

7. Animal species differ in the extent to which hormones exert a permanent organizational effect on behavior or on gonadotropin production. Specifically, sexual behavior in primates appears to be less dependent on gonadal hormones than in rodents; for example, when estrogens are given in appropriate amounts to male rhesus monkeys of any age a positive release of luteinizing hormone (similar to that of the normal ovulatory surge in females) can be induced (Resko, 1975). This suggests that hormones play a less permanent role in mediating function of the central nervous system in the primate.

8. Even when hormones are involved in mediating specific aspects of behavior, stereotyping can also play a role. For example, development of the normal male song pattern in some bird species requires both the action of androgen and exposure of the developing male to a mature male of the same species (Arnold, 1980). In this case the hormone acts both directly and by influencing learning patterns.

In summary, the role of gonadal steroids in sexual behavior involves development and function of the central nervous system, the development of the genital tract in the two sexes, reflex, sensory, and motor aspects of neurosensory arcs, and the integration of the various neural subsystems that constitute the behavioral process.

VI. CONTROL OF LIBIDO AND POTENTIA

The investigation of behavior patterns in the human is more complex at every level, especially the attempt to define, quantify, and elucidate the origins of behavior. For the purposes of this discussion we shall use the terms libido for the instinctual sexual drive and potentia for the ability to perform and complete sexual intercourse. These functions are not generally considered to be sexually dimorphic, but to a certain extent they are controlled by gonadal hormones. The simplest question is whether copulation is possible in humans in the absence of gonadal steroids. In male animals, orchidectomy is followed by retention of mating capacity for a variable period followed by eventual failure, whereas ovariectomy causes complete and immediate abolition of female mating behavior (Davidson, 1972). In the human, prepubertal castration appears to prevent uniformly the development of normal male behavior, and orchidectomy in the adult has sequelae similar to those in animals, i.e., castration of adult men causes a decline in sexual behavior with only occasional castrated subjects continuing to have intercourse over a period of years (Beach, 1948; Bremer, 1959). Furthermore, physiological androgen replacement therapy in such patients causes a rapid and reliable restoration of male sexual activity (Davidson et al., 1979). Thus, there is a considerable similarity between the hormonal control of male sexual behavior in man and animals.

However, removal of ovarian secretions by oophorectomy or via the natural menopause does not have a consistent deleterious effect on sexual activity in women, in dramatic contrast to the situation in female animals (Davidson, 1972). The usual interpretation is that once sexual patterns are fixed in women, sexual drive must be endocrine independent. This interpretation may be incorrect since removal of the human ovaries does not totally ablate the production either of testosterone or of estrogen, whereas adrenalectomy (Waxenberg et al., 1959) and hypophysectomy (Schon and Sutherland, 1960) do cause a profound decrease in sexual desire in castrated women. Consequently, it is possible that the sexual life of the human female is as hormone dependent as that of animals. Adrenal androgen (ablated by either hypophysectomy or adrenalectomy) could have a direct effect on sexual desire in women, or adrenal androgen acting as prohormone for estrogen synthesis in peripheral tissues could supply sufficient estrogen for maintenance of sexual drive in the absence of the ovaries. (The issue is even more uncertain as to whether hormones are involved in the genesis of normal sexual drive at female puberty.)

A similar problem of interpretation exists as to exactly which hormones regulate male sexual behavior. Occasional castrate males of all species sustain a capacity and drive for intercourse for long periods (Davidson, 1972; Bremer, 1959). In the castrate male considerable estrogen and small amounts of testosterone are formed in peripheral tissues from adrenal androgen (Siiteri and MacDonald, 1973), and in some animal species estradiol enhances the effect of androgen on male sexual drive (Wilson, 1975). Thus, the small amounts of testosterone and/or estrogens formed via this mechanism may be enough to sustain libido and potentia in some adult male castrates. In other words, libido and potentia would be preserved only in those men able to produce sufficient hormones by the adrenals and/or peripheral tissues.

In summary, testicular secretions play an important role in the male sexual drive of humans and animals, whereas ovarian secretions appear to be more important in influencing the female sexual behavior of animals than of women. One effect of this type of analysis has been to cast doubt on the applicability of animal models for the understanding of human reproductive behavior.

Although indirect, a second type of evidence also raises doubts about the validity of animal models for the analysis of human behavior; namely, there does not appear to be a permanent imprinting by steroid hormones on the pattern of gonadotropin secretion in the human. [In this regard, the human is similar to the rhesus monkey (Resko, 1975).] For example, the normal tonic pattern of gonadotropin secretion in adult men can be altered to the cyclic pattern characteristic of the female by administration to men of ovarian steroids so as to mimic the pattern of secretion of the normal ovary (Kulin and Reiter, 1976; Barbarino and de Marinis, 1980). This suggests that gonadal hormones have no organizational effect on central nervous system function in the human since the male pattern can be reversed. This finding, together with the observation that gonadal steroids are not required to maintain sexual libido in women, suggests that the effects of hormones on central nervous system functions—both behavioral and nonbehavioral—are different in the human than in lower animals.

VII. SEXUAL IDENTIFICATION IN THE HUMAN

In contrast to sexual drive—which is not usually considered to be sexually dimorphic—aspects of sexual behavior relating to identification are fundamentally different in males and females. For the present purposes the two important concepts are gender identity and gender role. Gender identity is defined as the unified and persistent experience of oneself as male, female, or ambivalent. Gender role is composed of the actions, activities, and behavior that indicate to others the degree to which one is male, female, or ambivalent. The factors that constitute gender role are obviously influenced by a variety of cultural and social variables, since actions and activities of the two sexes vary in different societies. Consequently, the patterns by which one identifies oneself as male or female vary. Two other aspects of dimorphic behavior are beyond the scope of the present discussion: sexual orientation (whether sexual object choice is heterosexual, homosexual, or bisexual) and parenting (the desire and capacity for the care of children). Knowledge of endocrine influences on both processes is sparse (for review, see Baker, 1980).

Vagueness in the definition and understanding of gender identity and gender role reflects the fact that it is difficult to quantify these parameters in any meaningful way and more difficult to devise means of investigating their provenance. Appropriately controlled experiments that would allow rigorous identification of the determinants of sexual identification and behavior cannot be performed in human beings.

As a consequence, a major emphasis in the study of human sexual behavior has been the analysis of gender role and behavior in subjects with histories of endocrine abnormalities, particularly studies of patients with abnormalities of sexual development. To understand the limitations in the usefulness of these pathological states for the analysis of human behavior it is necessary first to consider briefly the pathophysiology of abnormal sexual development.

A. Abnormal Sexual Development

Derangements of any of the three primary processes involved in sexual differentiation can cause abnormal sexual development. Disorders of chromosomal, gonadal, and phenotypic sex have all been recognized in the human (Table I). Such abnormalities can arise from environmental insult, as in the ingestion of a virilizing drug; aberrations of the sex chromosomes, as in 45,X gonadal dysgenesis; developmental birth defects of multifactoral etiology, as in most cases of hypospadias; or hereditary disorders resulting from single-gene mutations, as in the testicular feminization syndrome. The pathogenesis, clinical spectrum, endocrine pathology, and functional derangements that accompany these disorders have been reviewed extensively in the recent past and will not be considered here (Wilson and Walsh, 1979; Wilson and Goldstein, 1975). However, several aspects of abnormal sexual development are relevant to the analysis of human sexual behavior.

TABLE I

Occurrence of Ambiguous Genitalia and Changes in Gender Role in Abnormalities of Sexual Development

First, there is considerable variation in the seriousness of the phenotypic defects that eventuate in the various types of abnormal sexual development. For example, men with the 47,XXY Klinefelter syndrome or with the XX male syndrome develop as normal men (albeit infertile because of azoospermia) and express endocrine abnormality only later in life. Likewise, subjects with 45,X gonadal dysgenesis or the syndrome of pure gonadal dysgenesis develop a female phenotype, and most patients with true hermaphroditism have unambiguous male or female phenotypes. Thus, many patients with abnormalities of sexual development end up nevertheless with unequivocal male or female phenotypes. This is the consequence either of the fact that the formation of testicular hormones was sufficient to induce a male phenotype or that the failure of production of testicular hormones is complete enough to result in formation of a female phenotype. If hormones are involved in the formation of gender identity, in most patients the hormonal tendency would correspond with the anatomical development and hence with the sex assignment at birth.

Second, disorders that appear phenotypically similar can result from very different mechanisms. For example, some patients with errors in chromosomal sex, such as mixed gonadal dysgenesis, have phenotypes similar to those of patients with abnormalities of phenotypic sex, such as 5α-reductase deficiency. Since different disorders differ in their pathophysiology, it is essential that the diagnosis be established before any valid interpretation can be drawn as to behavioral consequences of a given pathology.

Third, ambiguity of genital development (and hence confusion as to appropriate gender assignment at birth or in subsequent life) occurs in only a few disorders: (1) The testes do not produce sufficient hormones to virilize the male embryo—either because of developmental abnormality of the testes (as in mixed gonadal dysgenesis) or because of a hereditary defect in one of the enzymes required for testosterone biosynthesis. (2) Sufficient testosterone is synthesized by the testes, but due to an inherited abnormality that affects the molecular machinery of androgen action (frequently the receptor protein or 5α-reductase enzyme) the hormone cannot act to virilize the embryo normally. (3) Overproduction of androgen occurs in the female embryo as in congenital adrenal hyperplasia due to deficiency of the 21-hydroxylase enzyme. (Bongiovanni, 1978). If hormones are involved directly or indirectly in development of gender identity one would predict that gender identity would be most imperfect in patients with ambiguous genitalia. However, even if this were true, gender identity would not be expected to be influenced in every patient. All defects that cause ambiguous genitalia vary in severity among affected individuals and consequently result in variable degrees of abnormal genitalia. For example, the external phenotypes of chromosomal males with abnormalities of the androgen receptor or of chromosomal females with 21-hydroxylase deficiency can span an entire spectrum from male to ambiguous to female. One would not expect abnormalities of gender identity in those individuals with minor or no defects in genital development.

Fourth, even when the degree of ambiguity of the external genitalia is similar, disorders can have different times of onset and different long-term endocrine consequences. For example, disorders of androgen synthesis and/or action influence embryonic development beginning at the end of the first trimester (Wilson, 1978; Griffin and Wilson, 1980), whereas adrenal function—and hence adrenal virilization in 21-hydroxylase deficiency—does not begin until later in embryogenesis (Bongiovanni, 1978). Furthermore, adult males with 17β-hydroxysteroid dehydrogenase deficiency, mixed gonadal dysgenesis, or 5α-reductase deficiency may have a normal endocrine profile for a postpubertal man despite the profound defect in androgen action during embryogenesis, whereas the testicular lesions in the Klinefelter syndrome and in the XX male become progressively severe so that plasma testosterone values, although initially normal, decline with age. Behavioral consequences might or might not occur in these disorders, depending on when in development gonadal steroids exert an effect on gender identity.

Thus, abnormalities of sexual development differ in their influence on the sexual phenotypes, their effects on hormone patterns at various times of life, their times of first manifestations during development, and their ultimate metabolic effects. Any interpretation as to possible behavioral consequences of a specific disorder has to take these various factors into account. That is, since various abnormalities have different effects on the anatomic and functional phenotypes, different behavioral consequences would be predicted in various disorders if hormones are involved in the genesis of human sexual behavior. For these reasons, it is necessary to be especially cautious in interpreting any negative result, specifically one that fails to support an effect of hormones on gender development.

B. Behavioral Studies in Patients with Abnormalities of Sexual Development

While different forms of abnormal sexual development have been lumped together in many behavioral studies, detailed studies have been performed in six groups of patients with specific diagnoses:

1. Females exposed to excess androgens as a result of the syndrome of congenital adrenal hyperplasia develop a variable degree of virilization of the external genitalia. Gender identity is usually female despite the presence of virilization and despite the fact that discernible effects can be delineated in certain aspects of gender role behavior, generally a tomboyishness and characteristic male energy expenditure (Baker, 1980; Money et al., 1955; Ehrhardt et al., 1968a,b; Money, 1969).

2. Children who were exposed to exogenous estrogen or progesterone during gestation usually have appropriate male or female phenotypes. These agents, at best, have minor effects on sexually dimorphic behavior and no discernible effect on gender identity (Yalom et al., 1973; Reinisch, 1974, 1976, 1977; Meyer-Bahlburg et al., 1977; Reinisch and Karow 1977; Ehrhardt et al., 1977).

3. Males with complete androgen resistance and the syndrome of testicular feminization develop a female phenotype. Such patients have testes and male testosterone levels but cannot respond to androgen and consequently differentiate as phenotypic women. Gender identity and gender role in such patients are unequivocally female in accord with anatomy and sex assignment, and such patients rank high in all femininity quotients (Masica et al.,