The Role of Androgens in Male Gender Role Behavior

0163-769X/99/$03.00/0
Endocrine Reviews 20(5): 726 –737
Copyright © 1999 by The Endocrine Society
Printed in U.S.A.
The Role of Androgens in Male Gender Role Behavior
JEAN D. WILSON
Department of Internal Medicine, The University of Texas Southwestern Medical Center, Dallas, Texas
75235-8857
I. Introduction
mones in human behavior but, on the basis of studies of
II. Sexual Behavior of Animals
subjects with a variety of forms of human intersex and/or
III. Control of Libido and Potentia in Humans
endocrine abnormalities, it has been the predominant view
IV. Gender Identity/Role Behavior in the Human
that human behavior is more complex than that of other
A. Normal and abnormal sexual development
species and that human gender identity and gender role
B. Behavior studies in subjects with abnormal sexual
behavior are determined primarily, if not exclusively, by
development
psychological and social forces (reviewed in Ref. 2). Accord-
V. Gender Role Behavior in Individuals with Male
ing to this anthropocentric view, the human species has been
Pseudohermaphroditism
emancipated from biological controls so that the hormones
A. 17 -Hydroxysteroid dehydrogenase 3 deficiency
that mediate this aspect of sexual behavior in animals do not
B. Steroid 5 -reductase 2 deficiency
play a significant role in controlling human behavior (3). As
C. Features common to 17 -hydroxysteroid dehydro-
summarized by Herdt (4),“the sex of rearing outweighs the
genase 3 and steroid 5 -reductase 2 deficiencies
biological sex in the development of gender identity and
D. Androgen receptor mutations
social identity.”
VI. Discussion
This belief that hormones do not play a significant role in
VII. Conclusions
controlling human gender role behavior persists despite a
large body of evidence to the contrary, indicating that an-
drogens play an important role in human male gender iden-
I. Introduction
tity/behavior. This evidence stems largely from the work of
INMOSTspeciesallaspectsofreproductionarecontrolled Imperato-McGinley and her colleagues (5, 6), who docu-
by hormones secreted by the ovaries and testes. Such
mented that genetic males with either of two autosomal
functions include the formation of the sexual phenotypes
recessive mutations that impair androgen synthesis or an-
during embryogenesis, sexual maturation at the time of pu-
drogen metabolism during embryogenesis, and hence cause
berty, and various forms of sexual behavior including sex
formation of female external genitalia and female sex of
drive and potentia, gender-typical behavior, and, in some
rearing in genetic males, may change gender role behavior to
species, traits such as aggression, the drive for dominance,
male at or after the time of expected puberty. The fact that
and parenting behavior (reviewed in Ref. 1).
single gene mutations can be associated with change in gen-
In humans gonadal steroids are responsible for phenotypic
der role behavior raises fundamental questions about the
sexual differentiation, sexual maturation, and development
factors that regulate human sexual behavior.
of libido and potentia. Human sexual behavior also involves
The molecular biology of these two autosomal recessive
gender identity, the perception of oneself as male or female,
disorders has been explored in some detail. The cDNAs and
and gender role behavior (also termed social sex or social
genes that encode the two critical enzymes involved, 17 -
identity), the various processes by which gender identity is
hydroxysteroid dehydrogenase 3 and steroid 5 -reductase 2,
communicated to others. Gender identity cannot be assessed
have been cloned, and a great deal has been learned about
in animals, and gender role behavior in animals can be dif-
the underlying pathophysiology. This review is designed to
ficult to separate from sexual orientation. Whether gonadal
consider some of the implications of these studies for un-
steroids are involved in the development of human gender
derstanding human behavior.
identity and role behavior is difficult to examine. These two
aspects of behavior are normally in accord, but most studies
on this subject focus on gender role behavior because the
II. Sexual Behavior of Animals
change of legal registration of sex from one gender to another
is unambiguous, whereas gender identity can be a graded
The role of gonadal hormones in animal behavior has been
character and difficult to quantify. It is obviously not possible
the subject of several extensive reviews (7–12). For the pur-
to devise definitive experiments to examine the role of hor-
poses of this discussion certain aspects of this relationship
deserve emphasis:
1. Sexually dimorphic behaviors of a variety of types are
Address reprint requests to: Dr. Jean D. Wilson, Department of In-
regulated by gonadal steroids, including the songs and mat-
ternal Medicine, UT Southwestern Medical Center, Room J6 –110, 5323
Harry Hines Boulevard, Dallas, Texas 75235-8857 USA. E-mail: jwils1@
ing behaviors of birds, copulatory patterns in mammals, and
mednet.swmed.edu
complex forms of ritual behavior such as musth in elephants
726

---

October, 1999
GENDER ROLE BEHAVIOR
727
and male dominance in mice. By way of illustration, male
organizational effects. In particular, organizational effects
and female rodents differ in the types of sexual postures they
appear to be less clear cut in primates than in rodents (22);
assume during coitus; these behaviors can be changed to
for example, the administration of estrogens in appropriate
those of the other sex by appropriate hormonal manipula-
amounts to male rhesus monkeys of any age elicits a positive
tion.
release of LH, analogous to the ovulatory surge of LH release
2. Androgens and estrogens are formed in both males and
in females (7).
females, and both hormones may play a role in the physi-
7. Even when hormones are involved in specific aspects of
ology of both sexes. However, androgens (and androgen
behavior, stereotyping can also play a critical role. For ex-
metabolites including estrogens in some species) are the pri-
ample, development of the characteristic male song pattern
mary determinants of male sexual behavior (13).
in bird species such as the zebra finch and canary require
3. Gonadal steroids act in the central nervous system by the
both the action of androgen in the central nervous system and
same receptor mechanisms that operate in peripheral tissues.
exposure of the immature male to a mature male of the same
Intracellular receptors for these hormones are expressed
species. Otherwise, the male will sing a garbled song instead
within specific regions of the brain (14, 15), and gonadal
of learning a song that will attract a female of the same
steroids may also exert central nervous system effects by
species (23). This androgen action is mediated by estrogenic
other mechanisms such as by influencing ion channels in cell
metabolites formed in the brain (24).
membranes (16 –18).
In summary, the role of gonadal steroids in sexual behav-
4. The behavioral effects of steroid hormones are due to
ior in animals involves, at a minimum, sexual dimorphism
interactions between peripheral and central actions of the
of the genital tracts, direct effects on the central nervous
hormones (2). One of the best studied paradigms of sexual
system, sensory and motor aspects of neurosensory reflexes,
behavior in the mammal is the mounting reflex of the female
and, probably, integration of the various neural subsystems
rat. Mounting of a female rat in estrus by a male causes the
that control the behavioral process.
female to extend the hind legs and elevate the rump, thus
dorsiflexing the vertebral column. These actions require sen-
sory input from the rump and involve a well defined neural
III. Control of Libido and Potentia in Humans
reflex that includes motor and sensory components and ste-
roid-mediated effects in the central nervous system. While
For the purposes of this discussion the term libido refers
there is no doubt that the central nervous system plays a vital
to the instinctive sexual drive, and potentia refers to the
role in the hormonal control of sexual behavior, different
ability to perform and complete sexual intercourse. These
behaviors may be influenced to different degrees by central
functions are not considered to be sexually dimorphic, but
and peripheral actions of the hormones. Even under defined
they are influenced by gonadal hormones. In animals cop-
laboratory conditions, it may be difficult to quantify the
ulation does not occur in the absence of gonadal hormones.
relative contribution of each to a given action (2).
In the males of most species, mating capacity is maintained
5. In the rodent the surge of testosterone secretion during
for a limited period after castration, followed by progressive
the neonatal period appears to play a vital role in virilizing
failure, and ovariectomy of female animals causes immediate
hypothalamic function, e.g., in imprinting a tonic pattern of
cessation of female mating behavior (2). In the human, pre-
gonadotropin release in contrast to the cyclical secretory
pubertal castration of boys uniformly prevents the develop-
pattern in females. (Again, this action may be mediated by
ment of normal sex drive, and castration in the adult male
estrogenic metabolites of testosterone in the central nervous
produces sequelae similar to those in animals, i.e., a decline
system.) Whether the neonatal increase in testosterone levels
in sexual behavior with only occasional castrated men ca-
in the human male infant is of physiological significance is
pable of normal sexual activity after 2 yr (25, 26). Further-
not known, but blocking the neonatal surge delays the onset
more, physiological androgen replacement therapy in hy-
of puberty in male monkeys (19).
pogonadal men causes a rapid and predictable restoration of
6. Phoenix and colleagues (20) delineated two types of
male sexual drive (27, 28). Thus, the hormonal control of male
behavioral effects of steroid hormones. Organizational ef-
sexual behavior is similar in man and animals. The fact that
fects are exerted by hormones at a specific time in develop-
the administration of an aromatase inhibitor to testosterone-
ment; they appear to have permanent effects on function or
treated, castrated male monkeys impairs male sexual drive
behavior, effects that persist after the steroid is no longer
indicates that the estrogenic metabolites of testosterone play
present. Such organizational effects may be accompanied by
a critical role in the control of sex drive, (29), but studies of
changes in anatomical development of the brain (21). Acti-
the localization of radioactive steroid hormones in brain in-
vating effects require the continued presence of the steroid
dicate that some androgen actions in brain are mediated by
for full manifestation of the effects (20), e.g., the mounting
testosterone and/or dihydrotestosterone (30 –33).
response of the female rat during estrus. Although the de-
In contrast, removal of ovarian secretions by ovariectomy
lineation of these two types of behavioral effects is of con-
or via the natural menopause does not have a consistent
ceptual importance, there is considerable overlap between
effect on sexual activity in women (2). The common inter-
them. Organizational effects may be silent in the absence of
pretation is that once sexual patterns are fixed in women,
the proper hormonal signals, and concurrent phenomena
sexual drive is hormone independent. This interpretation
such as male copulatory behavior may persist for variable
may not be correct because removal of the ovaries does not
periods after castration. Furthermore, different animal spe-
impair formation of adrenal androgens. Adrenalectomy (34)
cies differ in the extent to which hormones exert permanent
or hypophysectomy (35) in previously castrated women is

---

728
WILSON
Vol. 20, No. 5
reported to decrease sexual desire. Consequently, it is pos-
A. Normal and abnormal sexual development
sible that the sexual life of women is as hormone-dependent
The embryos of both sexes develop in an identical fashion
as that of men. Adrenal androgen (which would be ablated
until the seventh week of gestation. Thereafter, the anatomic
by hypophysectomy or adrenalectomy) could have a direct
and physiological development in the two sexes diverge. As
effect on sexual desire in women or could act as a prohor-
formulated by Jost (39), normal sexual development in the
mone for the synthesis in extraglandular tissues of other
mammal depends on three sequential processes. The first
steroid hormones (36) that could maintain sexual drive in the
involves the establishment of genetic sex at the time of con-
absence of ovarian hormones. Whether hormones are in-
ception, the heterogametic sex (XY) being male and the ho-
volved in the genesis of normal sexual drive at female pu-
mogametic sex (XX) female. In the second phase information
berty is also unclear.
encoded on the sex chromosomes causes the establishment
A similar uncertainty exists as to whether adrenal steroids
of gonadal sex in which the indifferent gonad develops into
can affect male sexual behavior. Occasional castrate males of
either an ovary or a testis. The final stage involves the trans-
all species sustain a capacity and drive for intercourse for
lation of gonadal sex into phenotypic sex. In the presence of
long periods (2, 26). In the castrated human male, consider-
an ovary or in the absence of a functional gonad, the devel-
able estrogen and small amounts of testosterone are formed
opment of phenotypic sex proceeds along female lines. Mas-
in extraglandular tissues from adrenal androgens (37), and in
culinization of the urogenital tract and the external genitalia,
some animal species estradiol enhances the effect of andro-
in contrast, requires the actions of three hormones, antimul-
gen on male sexual drive (38). Thus, the small amounts of
lerian hormone, testosterone, and dihydrotestosterone, the
testosterone and/or estrogen formed from adrenal andro-
5 -reduced metabolite of testosterone. The formation of an-
gens may be enough to sustain libido and potentia in some
timullerian hormone in the fetal testis is essential for sup-
adult male castrates. In other words, libido and potentia
pression of the mullerian ducts and hence for prevention of
would be preserved in those castrated men able to produce
development of the uterus and fallopian tubes in the male.
sufficient active hormones by this mechanism.
Testosterone, which is synthesized primarily in the testes and
In summary, gonadal steroids play an important role in the
circulates in the plasma, converts the wolffian ducts into the
sexual drive of males of all species and in controlling the
epididymis, vasa deferentia, and seminal vesicles, and di-
sexual drive of female animals and possibly of women. Or-
hydrotestosterone, which is formed predominately in the
ganizational effects do not appear to play as important a role
target cells themselves, induces the formation of the male
in the control of gonadotropin secretion by gonadal steroids
urethra and prostate and the male external genitalia (Fig. 1).
in the primate as in lower animals. In brief, although there
Derangement of any of the three primary processes in-
may be slight differences, the control of libido and potentia
volved in sexual differentiation can cause abnormal sexual
appears to be similar in humans and animals.
development, resulting in disorders of chromosomal sex,
gonadal sex, or phenotypic sex. The pathogenesis, clinical
manifestations, endocrine pathology, and functional distur-
IV. Gender Identity/Role Behavior in the Human
bances that accompany these disorders have been reviewed
extensively and will not be considered here. However, sev-
In contrast to sexual drive, which is not sexually dimor-
eral aspects of abnormal sexual development are relevant to
phic, gender identity is fundamentally different in men and
the analysis of human sexual behavior.
women. Some of the ambiguities in the definition and un-
First, the phenotypic manifestations of the various abnor-
derstanding of gender identity and gender role behavior are
malities differ markedly. For example, men with 47,XXY
due to difficulties in quantifying these parameters and to the
Klinefelter syndrome or with the 46,XX male syndrome de-
fact that gender role behavior is influenced by cultural and
velop as men (albeit infertile) and have endocrine abnor-
social variables, as evidenced by the different actions and
malities only in later life. Likewise, women with 45,X go-
activities of the two sexes in different societies. Most studies
nadal dysgenesis or with 46,XX or 46,XY pure gonadal
of the subject have focused on social sex because the change
dysgenesis have a female phenotype, and most subjects with
of legal gender is an unequivocal event, but the net conse-
true hermaphroditism have unequivocal male or female phe-
quence may be to underestimate the real frequency of dis-
notypes. Thus, many if not most individuals with abnormal-
orders of gender identity because some individuals with
ities of sexual development end up with unambiguous male
discordant gender identity may not change gender role be-
or female anatomical development; this is the consequence
havior for personal reasons. It is also difficult to investigate
either of the fact that the formation of testicular hormones
the mechanisms that regulate gender identity/role behavior
was sufficient to induce a male phenotype or that the failure
because controlled studies of the process cannot be per-
of formation/action of testicular hormones was complete
formed in humans. As a consequence, a major emphasis in
enough to result in formation of a female phenotype. Since
the study of human sexual behavior has been the analysis of
sex assignment and the sex of rearing are determined by
gender role behavior in subjects with histories of endocrine
anatomical development, any direct hormonal effects on be-
abnormalities, particularly subjects with abnormalities of
havior in most individuals with abnormal sexual develop-
sexual development. To understand the limitations and use-
ment would not be apparent because they would correspond
fulness of studies of these pathological states for the analysis
to anatomical development and hence to gender assignment
of human behavior, it is necessary to consider briefly how
and sex of rearing.
such disorders arise.
Second, disorders that appear phenotypically similar can

---

October, 1999
GENDER ROLE BEHAVIOR
729
FIG. 1. Schematic diagram of testos-
terone biosynthesis in the Leydig cell of
the testis and of the mechanism of an-
drogen
action
within
target
cells.
17 HSD3, 17 -Hydroxysteroid dehy-
drogenase 3; 5 R2, 5 -reductase 2.
result from different mechanisms. For example, men with
pensatory mechanisms, adult males with 17 -hydroxy-
45,X/46,XY mixed gonadal dysgenesis can have phenotypes
steroid dehydrogenase 3 deficiency, mixed gonadal
similar to those of men with steroid 5 -reductase 2 deficiency
dysgenesis, or 5 -reductase 2 deficiency may have the en-
or with mutations of the androgen receptor. Since these dis-
docrine profiles of normal (or near normal) adult men despite
orders have distinct pathophysiologies, it is essential that
having profound defects in androgen action during embry-
diagnoses be unequivocally established before attempting to
ogenesis. In contrast, the endocrine defects in the Klinefelter
draw interpretations as to the behavioral consequences of
syndrome and in the 46,XX male become progressively more
any given abnormality.
severe with age. Any behavioral consequences of disorders
Third, ambiguity of genital development occurs in rela-
of sexual development would depend on when in develop-
tively few disorders of human intersex and is due to one of
ment gonadal steroids exert an effect on the behavior in
three mechanisms: 1) The testes do not produce sufficient
question.
hormones to virilize the male embryo— either because of
In summary, abnormalities of sexual development differ
developmental abnormality of the testes or because of a
in their effects on the sexual phenotypes, their effects on
defect in one of the enzymes required for testosterone bio-
hormone levels at various times of life, the times during life
synthesis; 2) Sufficient testosterone is synthesized by the
when they become manifest, and the ultimate metabolic con-
testes, but due to impairment of androgen action (usually a
sequences. Any interpretation as to possible behavioral con-
defect in the androgen receptor) the hormone cannot virilize
sequences of a specific disorder must take these various
the embryo normally; or 3) Overproduction of androgen
factors into account. Furthermore, since different abnormal-
occurs in the female embryo, as in congenital adrenal hy-
ities vary in the severity of their effects on the sexual phe-
perplasia due to deficiency of the steroid 21-hydroxylase
notypes and on endocrine function, some disorders would
enzyme. In these disorders gender assignment usually cor-
not be predicted to influence behavior even if hormones are
responds to the predominant or apparent anatomy. If hor-
normally involved in controlling the behavior in question.
mones are involved directly or indirectly in development of
For these reasons, it is necessary to be cautious in interpreting
gender identity, one would predict that gender identity/
negative results.
behavior would be more likely to be discordant or uncertain
in subjects with ambiguous genitalia. Nevertheless, all ab-
B. Behavioral studies in subjects with abnormal
normalities that cause ambiguous genitalia vary in severity
sexual development
among affected individuals and can cause variable pheno-
types. For example, the external phenotypes of males with
While different forms of abnormal sexual development
abnormalities of the androgen receptor and of females with
have been lumped together in some reports, sufficient num-
steroid 21-hydroxylase deficiency can span the entire spec-
bers of individuals with specific diagnoses have been studied
trum from male to ambiguous to female. One would not
to allow a few generalizations:
expect abnormalities of gender identity in those individuals
1. Exposure of females to excess androgens as a result of
with normal or near-normal genital development.
congenital adrenal hyperplasia causes a variable degree of
Fourth, even when the degree of ambiguity of the external
virilization of the external genitalia. Gender identity in such
genitalia is similar, disorders can have different times of
individuals is usually female even in virilized women and
onset and different long-term endocrine consequences. For
despite the fact that behavioral changes, such as tomboyish
example, disorders of androgen synthesis and/or action in-
behavior and characteristic male energy expenditure, have
fluence embryonic development beginning at about week 8
been described in some studies (40 – 47). [Occasional women
of gestation, whereas virilization in females with steroid
with congenital adrenal hyperplasia have male gender role
21-hydroxylase deficiency does not commence until some-
behavior, but this usually occurs in severely virilized women
what later in gestation. Furthermore, as the result of com-
in whom diagnosis and surgical correction of the external

---

730
WILSON
Vol. 20, No. 5
genitalia are delayed beyond infancy or in whom glucocor-
groups. In such a view, it is difficult or virtually impossible
ticoid therapy is inadequate (48, 49).]
in most studies of subjects with disorders of intersex to as-
2. Children exposed to exogenous estrogens or progesto-
certain the extent to which psychological/social and endo-
gens during gestation have appropriate male or female phe-
crine determinants contribute to this development because
notypes; in general, such agents have only minor effects on
the psychological/social forces almost always correspond
sexually dimorphic behavior and do not influence gender
with the anatomical and endocrine factors.
role behavior/identity (50 –56).
3. True hermaphrodites have both testes and ovaries (or
ovotestes) and may have male, female, or ambiguous phe-
V. Gender Role Behavior in Individuals with Male
notypes. In such individuals, gender role behavior usually
Pseudohermaphroditism
corresponds to the sex of rearing, although many of them
have anomalous secondary sexual characteristics (57).
Over the years occasional instances have been reported in
4. Women with gonadal dysgenesis have female pheno-
which individuals with abnormal sexual development have
types and female gender identity/gender role behavior (58).
undergone a reversal in gender role behavior (and presumed
Since such women are profoundly estrogen deficient, it has
reversal in gender identity) at some age after gender identity
been inferred that ovarian estrogen plays at best a minor role
is usually considered to be fixed irreversibly (reviewed in
in the evolution of female gender identity.
Ref. 65). The majority of these reports were published before
5. Men with the Klinefelter syndrome form sufficient an-
the means of making specific diagnoses as to the cause of the
drogen during embryogenesis to induce formation of a male
abnormal sexual development were widely available, and it
phenotype but usually have diminished androgen produc-
is not possible in retrospect to deduce the correct diagnosis
tion and enhanced estrogen production after puberty. Nev-
in many such reports, indeed even in some relatively recent
ertheless, most men with Kleinfelter syndrome have male
studies (66 – 69). Nevertheless, in analyzing these reports two
gender role behavior, suggesting that these hormones play
conclusions seem justifiable: 1) Most such individuals are
no continuing role in gender identity/behavior at or after the
male pseudohermaphrodites with failure of virilization of
time of expected puberty (59).
the external genitalia and who were given a female sex as-
6. 46,XY women with profound androgen resistance due
signment at birth, and 2) The change in gender role behavior
to mutations of the androgen receptor develop a female
is usually from female to male. The fact that occasional in-
phenotype and unambiguous female behavior (see below)
dividuals with a disorder of human intersex change gender
(60 – 62).
role behavior long after the time of sex assignment was
The common thread in these various studies involving
clearly recognized by the anthropocentric school (63) and
many types of subjects and many different socioeconomic
was thought to result from childhood stigmatization of such
groups is that gender identity and gender role behavior
individuals because of their anatomical abnormalities (69).
usually develop in conformity with the sex assignment and
However, ambiguity of the genitalia cannot be the sole
the sex of rearing (62, 63). In other words, gender identity and
cause of changes in gender role behavior as illustrated by the
role behavior correspond with the predominant anatomical
case described by Stoller (70). This individual was thought to
development and hence with the prenatal hormonal milieu.
be a normal female at birth and was raised as a girl but
This conformity can withstand perturbations that include
exhibited tomboyish behavior from early childhood that be-
contradictory patterns in which girls virilize or boys feminize
came more and more masculine with time. She was an av-
during adolescence, tomboyish energy expenditure in girls,
erage student, but as adolescence ensued she became more
and incomplete development of the secondary sexual char-
and more withdrawn. Because of coarsening of the voice she
acteristics at puberty. Despite the inherent weaknesses in
was evaluated and found to be a genetic male with female
design in all such studies and despite the fact that none of the
external genitalia (including an apparently normal clitoris)
disorders constitutes a perfect experiment, the consistency of
but with testes in the labia majora. After psychiatric evalu-
the findings in such studies is impressive.
ation at age 14 she was told that she was a genetic male [the
The problem is that the findings are open to diametrically
diagnosis was subsequently established to be 17 -hydrox-
opposite interpretations. The predominant view—most elo-
ysteroid dehydrogenase 3 deficiency (5)]. She promptly
quently formulated by Money (63) and Lev-Ran (64)—is that
changed to male clothing and began to act, behave, and
sex assignment at birth influences parental attitudes and the
assume the role of a male. The parents decided to move to
manner in which infants are treated from the time of birth,
a new community; the boy’s grades improved, and he par-
and that these social factors are paramount in determining
ticipated in men’s sports in high school, obtained a university
human gender identity and gender role behavior, so pow-
degree in mathematics, and after urological surgery married.
erful as to be irreversible after early infancy. According to
This individual has been studied by several different groups
this view, any effects of hormones in influencing gender
over the years and apparently is a successful and well ad-
identity in the human are secondary and probably minor. A
justed man.
diametrically opposite interpretation is possible. Testicular
The fact that a single gene mutation could be associated
hormones could be important determinants of gender iden-
with a reversal of gender role behavior has far reaching
tity/behavior, but since they also control development of the
implications for understanding gender behavior, and in the
external genitalia (and hence determine sex assignment and
ensuing years it has been established that female-to-male
the sex of rearing) gender identity and anatomical sex would
reversal of gender role behavior appears to be a common
almost invariably be the same in these various patient
feature of two autosomal recessive forms of male pseudoher-

---

October, 1999
GENDER ROLE BEHAVIOR
731
A characteristic feature of the disorder is that the defect in
virilization (and the abnormality in testosterone levels) be-
comes less severe with time, and many affected individuals
eventually have near-normal male plasma testosterone levels
(78). Testosterone in these individuals can be formed by two
mechanisms. Namely, some mutant enzymes are neverthe-
less capable of some testosterone synthesis when LH and
FIG. 2. The 17 -hydroxysteroid dehydrogenase reaction for the in-
androstenedione levels are high, whereas individuals with
terconversion of androstenedione and testosterone. Androstenedi-
more severe mutations appear to convert androstenedione to
one is believed to be converted to testosterone by isoenzymes 3 and
testosterone in extraglandular tissues by the action of one or
5, and testosterone can be oxidized to androstenedione by isoen-
more of the unaffected isoenzymes, probably isoenzyme 5
zymes 2 and 4.
(78). The consequence is that an alternate pathway for tes-
maphroditism—5 -reductase 2 deficiency (6, 71) and 17 -
tosterone formation is present in all patients and that tes-
hydroxysteroid dehydrogenase 3 deficiency (5, 72, 73) (Fig.
tosterone formed in this way can cause considerable viril-
1). A similar change in gender role behavior has been de-
ization after the time of expected puberty.
scribed in genetic males with 3 -hydroxysteroid dehydro-
This disorder is rare and believed to be even less common
genase deficiency (74), an even rarer autosomal recessive
than 5 -reductase 2 deficiency. Andersson and colleagues
form of male pseudohermaphroditism, and in a few indi-
(76, 78, 79) have identified 16 different mutations in affected
viduals with mixed gonadal dysgenesis (65). This review
subjects that cause 12 different amino acid substitutions, 3
focuses on the two more common disorders, and we will
splice junction abnormalities, and 1 small deletion that
compare the consequences of mutations in these two en-
causes a frame shift. The latter types of mutations are be-
zymes with those of mutations of the androgen receptor on
lieved to preclude the formation of functional enzyme, but
gender role behavior.
the missense mutations impair enzyme function to variable
degrees (78, 79).
In addition to the Stoller patient, several individuals iden-
A. 17 -Hydroxysteroid dehydrogenase 3 deficiency
tified and raised as females have undergone a changed gen-
The 17 -hydroxysteroid dehydrogenase reaction is the ter-
der role behavior from female to male at the time of expected
minal step in the synthesis of testosterone in the Leydig cell
puberty (72, 73, 76, 80). In some case reports affected indi-
and of estradiol in the granulosa cell, and the rate of the back
viduals were too young to assess gender identity, and a few
reaction in extraglandular tissues plays a major role in de-
affected subjects have been raised from the beginning as
termining the steady state levels of these steroids in tissues
male. However, in a number of families, affected adult in-
(Fig. 2). Isoenzymes that perform these reactions are encoded
dividuals have female sexual identity/role behavior (75, 78).
by at least five genes (75) (Table 1), and mutations of the type
If one excludes case reports in infants and small children,
3 isoenzyme (76) are responsible for a rare, autosomal re-
gender role reversal appears to occur in about half of affected
cessive form of male pseudohermaphroditism originally de-
males. Because change in gender role behavior is so common
scribed by Saez and colleagues in 1971 (77). The typical
in this disorder, careful psychiatric evaluation must be ob-
features of this disorder are summarized in Table 2. In brief,
tained before any corrective surgery is undertaken. Although
affected 46,XY infants have female external genitalia, despite
it is not certain why this behavioral change occurs only in
the presence of testes and male wolffian structures; they are
some patients, this difference is not due to variations in the
usually assigned a female gender at birth and raised as fe-
severity of the mutation. Changes in gender role behavior
males. They usually come to medical attention because of
have occurred in one individual who is believed to make no
virilization at puberty or because of failure to menstruate. On
functional isoenzyme 3 as a result of a splice junction defect
endocrine evaluation they have low testosterone levels (for
(72, 76) and in the Arab family from Gaza who make a
men), normal ratios of plasma testosterone to dihydrotest-
kinetically abnormal enzyme that nevertheless can function
osterone, and variable estrogen levels. The diagnosis is made
partially (73, 76). While affected males from the Gaza family
by finding androstenedione levels that are usually 10 times
usually change gender role behavior from female to male, it
normal [Stoller’s patient had typical endocrine features for
is interesting that two Brazilian sisters with the identical
this disorder (5).]
mutation (R80Q homozygotes) have female gender role be-
TABLE 1. Comparison of human 17 -hydroxysteroid dehydrogenase isoenzymes
Isoenzyme
1
2
3
4
5
Size (amino acids)
327
387
310
737
323
Chromosome location of gene
17q21
16q24
9q22
5q2
10p14, 15
Tissue expression
Ovary, placenta
Endometrium, placenta, liver
Testis
Ubiquitous
Liver, skeletal muscle
Subcellular localization
Cytosol
Microsomes
Microsomes
Peroxisomes
Cytosol
Substrate preference
C steroids
C , C , C
steroids
C , C
steroids
C
steroids
C , C
steroids
18
18
19
21
18
19
18
19
21
Preferred cofactor
NADPH
NAD
NADPH
NAD
NADPH
Catalytic preference
Reduction
Oxidation
Reduction
Oxidation
Reduction
Activity in 17 HSD deficiency
Normal
Normal
Impaired
—
—

---

732
WILSON
Vol. 20, No. 5
TABLE 2. 17 -Hydroxysteroid dehydrogenase 3 deficiency
TABLE 3. Comparison of human 5 -reductase isoenzymes
Inheritance
Autosomal recessive
Isoenzyme
Phenotype
Males
1
2
Male Wolffian structures
Female urogenital sinus and external
Size (amino acids)
259
254
genitalia
pH Optimum
Neutral to basic
Acidic
(Females asymptomatic)
Chromosome location
5p15
2p23
Hormone profile
Low testosterone levels
of gene
High androstenedione levels
Gene organization
5 Exons/4 introns
5 Exons/4 introns
Low or normal estrogen levels
Expression in prostate
Low
High
Normal testosterone/dihydrotestosterone
Activity in 5 -reductase
Normal
Impaired
ratios
deficiency
Gender assignment
Female
at birth
osterone. The measurable plasma dihydrotestosterone (and
the subsequent partial virilization at puberty) can arise by
havior (76). Furthermore, in at least one family with another
two mechanisms; in individuals with mild kinetic abnor-
mutation (A203V), one affected individual changed gender
malities of enzyme function some dihydrotestosterone may
role behavior to male whereas the other is a married female
be derived from the mutant enzyme 2, whereas in individ-
(76).
uals with mutations that prevent formation of a functional
enzyme 2 plasma dihydrotestosterone can be derived from
B. Steroid 5 -reductase 2 deficiency
enzyme 1 (82). It is of interest in this regard that the activity
The conversion of testosterone to dihydrotestosterone
of enzyme 1, the principal isoenzyme in nongenital skin, is
(Fig. 3) changes a weak hormone to a more potent hormone
initially low and increases at the time of expected puberty
and is essential for many androgen actions (reviewed in Ref.
(84), probably explaining why impairment of virilization in
81). This reaction is irreversible and is mediated by two
these subjects is more complete during embryogenesis than
enzymes that are encoded by separate genes (Table 3). En-
at the time of expected puberty.
zyme 2 is the principal enzyme in the male urogenital tract
Imperato-McGinley et al. (6) reported that 18 of 19 affected
and plays a critical role in the virilization of the external
individuals from one family with 5 -reductase deficiency in
genitalia and urogenital sinus during embryogenesis. En-
the Dominican Republic were initially raised as females but
zyme 1, which after puberty is expressed in many tissues,
subsequently changed gender role behavior to male at the
may play a role in androgen metabolism in sebaceous glands
time of expected puberty. A similar phenomenon has been
and in the central nervous system.
described in other parts of the world: about two-thirds of
5 -Reductase deficiency causes an autosomal recessive
individuals raised as females change to male gender role
form of male pseudohermaphroditism in which the pheno-
after the time of expected puberty (82). In one study of 16
type resembles that in 17 -hydroxysteroid dehydrogenase 3
patients from 10 families studied by the same psychologist,
deficiency. Namely, virilization of the external genitalia is
3 individuals retained a female gender role, 12 changed to
impaired, and affected males are usually assigned a female
male gender role, and one was raised as a male (85), and in
gender at birth and raised as females (the mutation appears
a study of 10 affected individuals from 8 families studied in
to be silent in women) (Table 4). When the cDNAs for these
another unit, 6 changed gender role behavior to male, 3 have
genes were cloned, it was found as expected that the muta-
female gender role behavior, and 1 was raised as a male (86,
tions involve the gene for enzyme 2 (reviewed in Ref. 81), and
87). Thus, reversal of gender role behavior may be even more
45 different mutations have been described to date, including
common in this disorder than in 17 -hydroxysteroid dehy-
35 different missense mutations, 3 premature stop codons, 3
drogenase 3 deficiency. As in 17 -hydroxysteroid dehydro-
small deletions and 1 deletion of the entire coding sequence,
genase deficiency, however, change in gender role behavior
1 small insertion, and a change from a stop codon to a
is not a simple function of the severity of the mutation, since
missense code (82, 83).
the phenomenon occurs with mutations that partially impair
As with 17 -hydroxysteroid dehydrogenase 3 deficiency,
the kinetics of the 5 -reductase and in at least one family with
these individuals virilize to a greater or lesser extent at the
a splice junction abnormality that is thought to prevent for-
time of expected puberty. They have normal male levels of
mation of functional enzyme (82). Furthermore, families
plasma testosterone and low (but not absent) dihydrotest-
have been reported in which some, but not all, affected in-
dividuals undergo the change in social sex (85, 88).
It is of interest that the earliest description of gender role
reversal and possibly of 5 -reductase deficiency appears to
be Herculine Barbin, a French woman who lived during the
19th century and who is believed to be the first person to have
changed legal sex from one gender to the other; her pheno-
type, including evidence from autopsy, is compatible with
the diagnosis (89, 90).
It should be emphasized that no prospective studies have
FIG. 3. The 5 -reductase reaction involved in the conversion of tes-
tosterone to dihydrotestosterone. Both isoenzymes 1 and 2 can per-
been done in either of these disorders so that it is not possible
form this conversion.
to be certain that gender identity before expected puberty

---

October, 1999
GENDER ROLE BEHAVIOR
733
TABLE 4. 5 -Reductase 2 deficiency
viduals to undergo considerable virilization. 5) Change in
gender role behavior in the two disorders at expected pu-
Inheritance
Autosomal recessive
berty is common but not universal; the reason for this in-
Phenotype
Males
consistency is not readily apparent and cannot be explained
Male Wolffian structures
Female urogenital sinus and
in any straightforward way by variations in the severity of
external genitalia
the mutations themselves. Whether this inconsistency might
(Females asymptomatic)
be explained by variability in the completeness of compen-
Hormone profile
Normal male testosterone levels
sation by the alternate pathways in the two disorders is
Normal estrogen levels
Decreased dihydrotestosterone levels
unknown.
Gender assignment
Female
at birth
D. Androgen receptor mutations
Although mutations that impair the function of the an-
TABLE 5. Features common to 5 -reductase 2 deficiency and 17 -
hydroxysteroid dehydrogenase 3 deficiency
drogen receptor (Fig. 1) can cause a phenotype that is similar
to those caused by the two enzyme deficiencies (Table 6),
1.
Impairment of virilization during embryogenesis is limited
gender role behavior in these subjects almost invariably cor-
to the external genitalia.
responds to the gender assignment at birth (83): if the im-
2.
46,XY males are given gender assignments and raised as
pairment of receptor function is severe enough at birth to
females.
3.
Considerable virilization takes place at the time of expected
cause the syndrome of complete testicular feminization and
puberty.
a female sex assignment, such individuals not only maintain
4.
In both disorders an alternate pathway exists so that the
the female sex assignment as adults but rank high in feminine
defects are incomplete, namely testosterone is formed in
traits as defined by psychological criteria (60, 61). Rare
17 HSD3 deficiency, and dihydrotestosterone is formed in
women with the syndrome of incomplete testicular femini-
5 R2 deficiency.
5.
Change in gender role behavior from female to male is
zation (whose mutated androgen receptors have partial re-
common but not universal.
sidual function and who virilize to a variable degree at pu-
berty) have been described in whom gender identity was
was ever unambiguously female. Indeed, several such per-
male despite being reared as female (92, 93); the significance
sons have stated in retrospect that they had been aware of
of this phenomenon is not clear. Many men with partial
uncertainties as to their correct gender from a very early age
androgen resistance and even less severe impairment of re-
(91); consequently, one cannot be certain that this is a change
ceptor function virilize sufficiently during embryogenesis to
in gender identity as contrasted to a resolution of a confused
result in a male sex assignment at birth and characteristically
gender identity— only that gender role behavior changes
have unequivocal male gender role behavior as adults (83).
from that of the sex of rearing to that of the genetic, gonadal,
The fact that complete testicular feminization is associated
and endocrine sex of the individual. This change could either
with a female gender role/identity despite the presence of
be the result of a change in gender identity or the resolution
testes and normal adult male levels of plasma testosterone
of an uncertain gender identity as virilization progresses at
indicates that any involvement of androgens in gender role
the time of expected puberty.
behavior must involve the androgen receptor. Furthermore,
since the extraglandular conversion of androgens to estro-
C. Features common to 17 -hydroxysteroid dehydrogenase 3
gens is normal in women with testicular feminization (Table
and steroid 5 -reductase 2 deficiencies
7) (37), the role of androgens in gender role behavior cannot
involve the conversion of androgens to estrogens, as appears
5 -Reductase 2 deficiency and 17 -hydroxysteroid dehy-
to be the case in some animal species (17, 23, 24). This con-
drogenase 3 deficiency share several common features (Table
clusion is supported by the fact that a man with a mutation
5): 1) In both, 46,XY males are given gender assignments at
that impaired function of the estrogen receptor (94) and two
birth; in this sense, gender role change, when it occurs, is a
men with profound aromatase deficiency (95, 96) have been
correction of a incorrectly assigned gender. 2) In both dis-
reported to have male gender identity.
orders the impairment of virilization during embryogenesis
is limited to the external genitalia; the internal urogenital
VI. Discussion
tract (testes, epididymis, vas deferens, seminal vesicle, and
ejaculatory ducts) is male in character, and the testes usually
What generalizations can be drawn about behavior from
descend into the inguinal canals or labia majora. 3) In both
the findings in these two single gene disorders? First, it seems
disorders considerable virilization takes place at the time of
TABLE 6. Androgen receptor mutations
expected puberty, particularly the growth of a phallus ca-
pable of erection; indeed, penile erections are the rule. 4) In
Inheritance
X-linked trait
both disorders an alternate pathway exists; testosterone can
Phenotype
Males
be formed by an alternate pathway in 17 -hydroxysteroid
Variable from women with testicular
dehydrogenase 3 deficiency, and dihydrotestosterone can be
feminization to undervirilized men
Hormone profile
Normal male testosterone and
formed by enzyme 1 in 5 -reductase 2 deficiency. Conse-
dihydrotestosterone levels
quently, in the postpubertal steady state in both conditions,
Increased estrogen production and levels
testosterone and dihydrotestosterone levels can be in the
Gender assignment
Varies with the anatomy
normal or near-normal male range, causing affected indi-
at birth

---

734
WILSON
Vol. 20, No. 5
TABLE 7. Estrogen formation in normal men and in subjects with
male pseudohermaphroditism
Estradiol
Estrone
Group
( g/day)
( g/day)
Normal men (4)
Total
45
60
Extraglandular aromatization
39
60
Secretion
6
0
Testicular feminization (4)
Total
77
114
Extraglandular aromatization
33
101
Secretion
44
13
Mean values have been taken from P. C. MacDonald et al. (37). The
number of individuals studied in each group is given in parentheses.
inescapable that androgen action is important for male gen-
der role behavior and probably for male gender identity as
well. This does not necessarily mean that androgens can
change gender identity/role behavior, only that they may
interfere with the development of a gender assignment not
FIG. 4. The phases of male sexual life as indicated by mean plasma
in accord with the genetic/endocrine sex. Second, this action
testosterone level as a function of age. Sperm production occurs only
cannot be mediated by the conversion of androgens to es-
during the adult phase. [Modified from Griffin JE, Wilson JD 1980
trogen; male gender identity appears to be normal in men
The testis. In: Metabolic Control and Disease, Bondy PK, Rosenberg
LE (eds) with permission from W.B. Saunders, Philadelphia. Based
with mutations of the estrogen receptor (94) or of aromatase
on the formulation of J.S.D. Winter et al.: J Clin Endocrinol Metab
(95, 96), and gender identity is female in 46,XY women with
42:679 – 686, 1976.]
testicular feminization despite normal or elevated plasma
estrogen levels (for men) and normal rates of extrogen for-
of potent inhibitors of both isoenzymes or double-knockout
mation by extraglandular aromatase (36, 37). Third, the an-
animals in which both 5 -reductase isoenzymes are missing.
drogen effect must be mediated by the androgen receptor
These model systems may make it possible to investigate the
since profound impairment of receptor function causes com-
effects of testosterone and dihydrotestosterone indepen-
plete testicular feminization that is characterized by female
dently.
gender identity/role behavior despite normal male levels of
No matter how important the implications of the findings
plasma testosterone (60, 61). It also follows that even partial
in these two disorders may be for understanding the control
androgen receptor function is usually adequate to support
of gender role and gender identity in the human, it is highly
male gender role behavior, since most men with mutations
unlikely that abnormalities in androgen action are a common
that only partially impair androgen receptor function (Rei-
cause of transsexual behavior. Meyer et al. (97) studied 60
fenstein syndrome) have unequivocal male behavior even in
male-to-female transsexuals and 30 female-to-male transsex-
the presence of incomplete external virilization and consid-
uals; only two of these individuals (both female-to-male) had
erable feminization at the time of expected puberty.
an underlying endocrine abnormality so that, at best, less
This is not to say that there are not formidable unresolved
than a tenth of female-to-male transsexuals can be explained
aspects of this problem. For example, it is not known whether
by disordered action of androgen. In keeping with this con-
this action of androgen takes place during embryogenesis,
cept, Meyer-Bahlburg (98) argued convincingly that disor-
during infancy, or at the time of expected puberty, the phases
ders of gender identity in subjects with male pseudoher-
of male life associated with high levels of plasma testosterone
maphroditism are fundamentally different than gender
(Fig. 4). As stated above, several such individuals have re-
identity disorders in subjects that do not have a problem of
ported that they were conscious of gender conflicts from
human intersex in that the former group make the change in
early infancy (91), implying that the effect is either prenatal
gender role behavior with greater ease. Consequently, it is
or occurred during the neonatal period. Virilization at the
unlikely that studies of this type can provide insight into
time of expected puberty may influence this process but is
transsexualism per se, the etiology of which is believed to be
probably not critical because in some individuals [such as
outside the endocrine domain.
Stoller’s patient (70)], there is no evidence of genital ambi-
guity when the change in gender role behavior occurred.
VII. Conclusions
Likewise, in animal studies effects of androgens on behavior
can sometimes be identified in the absence of virilization of
Genetic and endocrine evidence indicates that androgen
the urogenital tract (10). It is also unclear whether the effect
action plays an important role in male gender role behavior;
of androgen on gender behavior is mediated at the level of
since gender identity and gender role behavior are normally
the central nervous system, the urogenital tract, or both; nor
in accord, androgen action is probably an equally important
is it intuitively clear how to investigate this issue in humans.
determinant of male gender identity. At the same time, it is
Finally, it is not known whether this androgen action is
also clear that androgen is not the sole determinant of these
mediated by testosterone or by dihydrotestosterone; insight
processes; the fact that many individuals with mutations of
into the latter question may be possible with the availability
the 5 -reductase and 17 -hydroxysteroid dehydrogenase

---

October, 1999
GENDER ROLE BEHAVIOR
735
enzymes do not undergo a change in gender role behavior
inhibition on estrogen receptor-like immunoreactivity in male rat
means that other factors–social, psychological, or biological–
brain. Neuroendocrinology 59:552–560
are of equal or greater importance in modulating human
16. McEwen BS 1980 Gonadal steroids: humoral modulators of nerve-
cell function. Mol Cell Endocrinol 18:151–164
sexual behavior. Indeed, the sex of rearing may be more
17. Celotti F, Massa R, Martini L 1979 Metabolism of sex steroids in the
important in this regard than the endocrine milieu under
central nervous system. In: DeGroot LJ (ed) Endocrinology, Grune
ordinary circumstances, and it may not be a coincidence that
and Stratton, New York, pp 48 –53
many (although not all) of the instances of reversal of gender
18. Wehling M 1997 Specific, nongenomic actions of steroid hormones.
Annu Rev Physiol 59:365–393
role behavior in these two disorders have occurred in coun-
19. Mann DR, Gould KG, Collins DC, Wallen K 1989 Blockade of
tries and/or ethnic groups in which men play a dominant
neonatal activation of the pituitary-testicular axis: effect on peripu-
role; in this situation, endocrine factors may be more impor-
bertal luteinizing hormone and testosterone secretion and on tes-
tant determinants of behavior than would be the case in more
ticular development in male monkeys. J Clin Endocrinol Metab
egalitarian societies.
68:600 – 667
20. Phoenix CH, Goy RW, Gerall AA, Young WC 1959 Organizing
Endocrine and psychological factors must interact to in-
action of prenatally administered testosterone propionate on the
fluence these behaviors. Perhaps the most appropriate ani-
tissues mediating mating behavior in the female guinea pig. Endo-
mal model for this aspect of human behavior is the song bird
crinology 65:600 – 667
in which androgen action in the central nervous system and
21. Gorski RA, Gordon JH, Shryne JE, Southam AM 1978 Evidence for
a pattern of behavior learned from a male of the same species
a morphological sex difference within the medial preoptic area of the
rat brain. Brain Res 148:333–346
are both necessary to learn a song that will attract a female
22. Reiter EO, Grumbach MM, Kaplan SL, Conte FA 1975 The response
of the same species (23). It may never be possible to assign
of pituitary gonadotropes to synthetic LRF in children with glu-
quantitative importance to the roles of the two processes in
cocorticoid-treated congenital adrenal hyperplasia: lack of effect of
human behavior, but it may be possible to determine how,
intrauterine and neonatal androgen excess. J Clin Endocrinol Metab
40:318 –325
where, and when in development androgen plays its role in
23. Arnold AP 1980 Sexual differences in the brain. Am Sci 80:165–173
this process.
24. Schlinger BA 1998 Sexual differentiation of avian brain and behav-
ior. Annu Rev Physiol 60:407– 429
25. Beach FA 1948 Hormones and Behavior. Harper (Hoeber), New
References
York, pp 20 –29
26. Bremer J 1959 Asexualization. A Follow-up Study of 244 Cases.
1. Phoenix CH, Goy RW, Young WC 1967 Sexual behavior: general
Macmillan, New York
aspects. In: Martini L, Ganong WF (eds) Neuroendocrinology. Ac-
27. Davidson JM, Camargo CA, Smith ER 1979 Effects of androgen on
ademic Press, New York, vol 2:163–196
sexual behavior in hypogonadal men. J Clin Endocrinol Metab 48:
2. Davidson JM 1972 Hormones and reproductive behavior. In: Levine
955–958
S (ed) Hormones and Behavior. Academic Press, New York, pp
28. Burris AS, Banks SM, Carter CS, Davidson JM, Sherins RJ 1992 A
63–103
long-term, prospective study of the physiologic and behavioral ef-
3. Money J 1994 The concept of gender identity disorder in childhood
fects of hormone replacement in untreated hypogonadal men. J
and adolescence after 39 years. J Sex Marital Ther 20:163–171
Androl 13:297–304
4. Herdt G 1997 Same Sex Different Cultures. Westview Press, Boulder,
29. Zumpe D, Bonsall RW, Michael RP 1993 Effects of the nonsteroidal
CO, p 47
aromatase inhibitor, fadrozole on the sexual behavior of male cyno-
5. Imperato-McGinley J, Peterson RE, Stoller R, Goodwin WE 1979
mologus monkeys (Macaca fascicularis). Horm Behav 27:200 –215
Male pseudohermaphroditism secondary to 17 -hydroxysteroid de-
30. Rees HD, Bonsall RW, Michael RP 1988 Localization and identi-
hydrogenase deficiency: gender role change with puberty. J Clin
fication of nuclear radioactivity in the pituitary gland and genital
Endocrinol Metab 49:391–395
tract after administration of 3H-testosterone, 3H-dihydrotestoster-
6. Imperato-McGinley J, Peterson RE, Gautier T, Sturla E 1979 An-
one, or 3H-estradiol to male rhesus monkeys. Cell Tissue Res 254:
drogens and the evolution of male-gender identity among male
139 –146
pseudohermaphrodites with 5 -reductase deficiency. N Engl J Med
31. Michael RP, Bonsall RW, Rees HD 1986 The nuclear accumulation
300:1233–1237
of [3H]testosterone and [3H]estradiol in the brain of the female
7. Resko JA 1975 Fetal hormones and their effect on the differentiation
primate: evidence for the aromatization hypothesis. Endocrinology
of the central nervous system in primates. Fed Proc 34:1650 –1655
118:1935–1944
8. Beach FA 1977 Courtship and mating. In: Beach FA (ed) Human
32. Michael RP, Bonsall RW, Rees HD 1987 Sites at which testosterone
Sexuality in Four Perspectives. Johns Hopkins Press, Baltimore, pp
may act as an estrogen in the brain of the male primate. Neuroen-
247–267
docrinology 46:511–521
9. Whalen RE 1977 Brain mechanisms controlling sexual behavior. In:
33. Michael RP, Rees HD, Bonsall RW 1989 Sites in the male primate
Beach FA (ed) Human Sexuality in Four Perspectives. Johns Hopkins
brain at which testosterone acts as an androgen. Brain Res 502:11–20
Press, Baltimore, pp 215–246
34. Waxenberg SE, Drellich MG, Sutherland AM 1959 The role of
10. Goy RW, Bercovitch FB, McBrair MC 1988 Behavioral masculin-
hormones in human behavior I. Changes in female sexuality after
ization is independent of genital masculinization in prenatally an-
adrenalectomy. J Clin Endocrinol Metab 19:193–202
drogenized female rheusus macaques. Horm Behav 22:552–571
35. Schon M, Sutherland AM 1960 The role of hormones in human
11. Wallen K 1996 Nature needs nurture: the interaction of hormonal
behavior. III. Changes in female