Download Logical Levels of Steroid Hormone Action in the

AMER. ZOOL., 21:233-242 (1981)
Logical Levels of Steroid Hormone Action in the
Control of Vertebrate Behavior1
ARTHUR P. ARNOLD
Department of Psychology and Brain Research Institute, University of California,
Los Angeles, California 90024
SYNOPSIS. Three steroid-sensitive neural systems are reviewed to suggest where hormones
act to modify neuroendocrine or behavioral functions: the system controlling the ovulatory
surge of luteinizing hormone in laboratory rats, the system controlling male copulatory
behavior in the rat, and the system controlling passerine bird vocalizations. In each, steroids act at several levels, including the final common path. This generalization is discussed
in light of some earlier conceptualizations of the levels of hormone action in behavioral
systems.
INTRODUCTION
herent picture of how internal and external stimuli affect instinctive behavior. One
class of internal events which he considered was hormonal influences. Tinbergen
summarized briefly the famous ethological
hierarchical model of centers which control instinctive behavior. The highest centers represented major instincts (e.g., reproduction), and such centers controlled
lower centers representing specific instincts such as fighting, courtship, locomotion, which in turn controlled lower
centers which coordinate specific movements involved in each behavior. In discussing where hormones might act in this
hierarchy, Tinbergen considered the early
work of Beach and others, and concluded
that "it is probable that hormones act exclusively on the higher centers" (p. 124).
He reached this conclusion because of the
specificity with which hormones activate
only certain behaviors. Since the animal
uses the same muscles for different instinctive behaviors (Tinbergen, 1951, p. 74) an
action of a hormone at a low level in the
hierarchy would seem to be incompatible
with the hormone's selective effect on the
behavior. In explaining the hierarchical
model, Tinbergen used the example of the
three-spined stickleback (Gasterosteus aculeatus). The lowest level of the hierarchy of
centers controlling reproductive behaviors
is comprised of the muscles involved in the
behaviors, e.g., swimming muscles. The
same muscles are used for both the zig-zag
1
From the Symposium on Social Signals—Compar- courtship swimming of the male stickleative and Endocrine Aspects presented at the Annual back, and for swimming during appetitive
Meeting of the American Society of Zoologists, 27- behaviors involved in food search. In this
Sex steroids have profound yet specific
effects on the behavior of vertebrates. For
example, in zebra finches (Poephila guttata)
castration reduces (and testosterone therapy increases) the frequency of courtship,
copulatory, and aggressive behaviors,
whereas other behaviors are relatively unaffected (Arnold, 1975). At what logical
levels do steroids act to bring about such
changes in behavior? For example, do they
act on sensory organs or afferents to the
nervous system, on motor pattern generators, on general arousal mechanisms, or
at other functional levels? The word logical is used here to imply that we wish to
determine the function of steroid target
tissues not only in terms of their physiology (e.g., primary sensory neurons) but
also in terms of their logical function within the network of tissues responsible for
the behavior. For example, a steroid hormone acting on a primary sensory neuron
could modify at least two logical functions:
it could alter the elicitation of the behavior
by sensory stimuli, or it could alter the
feedback of sensory information to a central pattern generator which is instrumental in assessment of the output of the generator.
In his classical treatise on mechanisms
and evolution of animal behavior, Tinbergen (1951) attempted to synthesize a co-
30 December 1979, at Tampa, Florida.
233
234
ARTHUR P. ARNOLD
instance, Tinbergen most likely would
have concluded that hormones affecting
courtship swimming induce a selective enhancement of certain patterns of muscle
activity rather than all contractions of the
swimming muscles, and these patterns
must be controlled at higher levels in the
nervous system.
Thirty years after Tinbergen published
this model in summary form, we still think
in terms of hierarchies of neural systems
involved in controlling behavior. This is
partly because the notion of hierarchy has
been common in the best described neural
systems involved in invertebrate behavior
patterns. But it is also because any diagram
of neural connections in a system underlying vertebrate behavior involves neurons
synapsing on others in a more or less linear
fashion (although loops, branches, or parallel pathways are possible) leading eventually to the motor effectors for the behavior. We have now reached the point
that, although we do not yet know the entire logical or functional circuitry for any
single hormone-sensitive vertebrate behavior pattern (which would be a prerequisite
for finding all of the sites where hormones
act to modify the behavior), we are familiar
enough with several neural systems to
make at least two conclusions about logical
levels of steroid hormone action. The first
conclusion is that sex steroids probably act
at many logical levels to modify neural
functioning, which leads to changes in behavior. Secondly, in some species, sex steroids act not only at logically high levels
(integrative levels rather high in the hierarchy, far removed from the motor effectors), but also at very low logical levels, in
the final common neural path and in the
muscles themselves. To draw these conclusions, I wish to examine three hormonesensitive neural systems.
pattern, yet there are functional aspects
which are similar to the two behavioral systems discussed below. In rats, there is an
increase in plasma estradiol levels on the
morning of proestrus which precedes release of luteinizing hormone releasing h o ^
mone (LHRH) from the median eminence
into the portal blood supply to the pituitary (Sarkar et al., 1976). Subsequently,
there is a large increase in LH secretion by
the pituitary which induces ovulation.
LHRH appears to have complex effects on
the pituitary during the LH surge. It not
only causes the release of LH, but also has
a self-priming effect. Exposure of the pituitary to LHRH increases its sensitivity to
subsequent LHRH, so that constant
amounts of LHRH are able to induce progressively greater releases of LH (Fink,
1979).
Estradiol is intimately implicated in triggering the ovulatory surge of LH. When
the proestrus increase in estradiol levels is
counteracted with antiestrogens, the surge
of LH does not occur, and simulation of
the proestrus rise in estradiol with exogenous estradiol injections will induce an LH
surge and ovulation (Brown-Grant, 1977).
There is evidence that estradiol exerts this
influence by acting at a number of sites
and by a variety of mechanisms. First of
all, it is clear that estradiol acts on the pituitary to increase the secretion of LH induced by a given amount of LHRH (Beck
et al., 1978; Fink, 1979). Estradiol is also
important for the LHRH self-priming effect on the pituitary, since the magnitude
of this LHRH-induced increase in pituitary sensitivity to LHRH is determined by
estradiol levels (Fink, 1979; Blake, 1978).
A second site of action of estradiol in
inducing the LH surge is in the medial
preoptic area (MPOA). This region has
been implicated consistently in the control
of ovulation. When the medial basal hyESTRADIOL CONTROL OF THE
pothalamus and pituitary are isolated from
OVULATORY LH SURGE
rostral afferents with knife cuts, ovulation
The first system is the neural network persists only if the rostral knife cut is ancontrolling the ovulatory surge of lutein- terior to the MPOA (Schwartz, 1972). Furizing hormone (LH) secretion by the an- thermore, electrical stimulation of the
terior pituitary of rats. This system is not MPOA is very effective in eliciting LHRH
a behavioral one since the endpoint is en- and LH secretion (Chiappa et al., 1977;
docrine secretion rather than a behavior Cramer and Barraclough, 1971). Since
LEVELS OF STEROID ACTION ON BEHAVIOR
MPOA neurons contain LHRH, it is possible that they are the secretory neurons
responsible for the rise in LHRH. However, the stimulation in MPOA may activate nerve fibers or cell bodies which syn#pse on others in a multisynaptic pathway,
which eventually fires LHRH-containing
neurons elsewhere, causing them to discharge LHRH. Significantly, estradiol increases the amount of LHRH and LH released by MPOA stimulation. A priori, it
could do this by acting on (a) afferents to
the MPOA (e.g., especially from the hippocampus, amygdala, or septum, which
contain estradiol receptors, Pfaff and Keiner, 1973); (b) on cells fired by stimulation
of the MPOA; (c) on neural elements
which receive direct or indirect inputs
from fibers or cells activated by MPOA
stimulation and which are involved in
LHRH release; and/or (d) on the pituitary
itself. Goodman (1978) has provided
strong evidence that the MPOA itself may
be a site of action, since implants of estradiol into the MPOA (but not into the medial basal hypothalamus) cause an LH
surge, even though both types of implants
raise pituitary estradiol levels by comparable amounts. This result fits nicely with
autoradiographic localization of estradiol
concentrating cells, since the MPOA contains many such cells (e.g., Pfaff and Keiner, 1973; Stumpf, 1968). There is also
evidence compatible with estradiol's action
on limbic regions afferent to the MPOA,
since if these afferents are sectioned, electrical stimulation of the MPOA is significantly less effective in triggering LH release (Chiappa et al., 1977). However, this
last finding does not necessarily imply an
action of estradiol outside of the MPOA
and hypothalamus, since the sensitivity of
MPOA neurons to estradiol could be dependent on their proper innervation.
Since stimulation of the hippocampus and
amygdala have been reported to affect
MPOA-stimulation induced LH release
and spontaneous ovulation (Kawakami et
al., 1973; Velasco and Taleisnik, 1969),
and because these regions contain cells
that accumulate estradiol (Pfaff and Keiner, 1973), one can also construct a plausible but speculative argument that estra-
235
diol acts in these regions to modulate the
release of LHRH which is controlled by
the MPOA.
In summary, it is likely that estradiol
triggers and modulates the ovulatory
surge of LH secretion by acting at three
(and possibly more) levels in the CNS-pituitary circuit: in the pituitary itself, in the
MPOA, and possibly in other limbic regions. The actions of estradiol in the
MPOA and pituitary may be partially redundant, since each has the effect of increasing LH secretion. These results also
suggest action of estradiol at the final common path of this circuit controlling LH secretion, i.e., in the pituitary itself.
In turning to studies of modulation of
behavior by steroids, we face some problems. The literature on neurohumoral
control of LH secretion is an order of magnitude larger than the literature on mechanisms of steroid control of any one vertebrate behavior pattern, so we must rely
on much less information. Secondly, LH
secretion varies in magnitude but not in
quality. This is not true for behavior. Each
behavior pattern is a complex and variable
sequence of muscular contractions so that
the steroid may change the behavior both
qualitatively and quantitatively. Therefore, precise measurement of the steroid's
effect has often been incomplete.
ANDROGEN EFFECTS ON MALE R A T
COPULATORY BEHAVIOR
The first behavioral system to be discussed here is masculine copulatory behavior of the laboratory rat. A large number
of CNS regions have been implicated in
the control or modulation of male copulatory behavior, including the MPOA, medial forebrain bundle, olfactory tubercle,
mammillary bodies, amygdala, hippocampus, neocortex, diencephalic-mesencephalic junction, and spinal cord (reviewed
by Malsbury and Pfaff, 1974; Hart, 1974;
Sachs, 1978; Larsson, 1979). The roles of
the amygdala, hippocampus, and neocortex appear to be minor, or modulatory,
since sexual behavior persists after lesioning these areas, although such lesions
cause deficits in this behavior. Yet the olfactory tubercle, amygdala, and hippocam-
236
ARTHUR P. ARNOLD
pus should be potential sites of action of
steroids, since these contain cells which accumulate estrogens and/or androgens
(e.g., Pfaff and Keiner, 1973; McEwen,
1976; Stumpf and Sar, 1978). However,
steroid effects on these regions which influence male sex behavior have apparently
not been investigated extensively. Good
evidence exists, however, for steroid actions in the MPOA to facilitate mounting,
intromission, and ejaculation, since lesions
of this region eliminate these male sexual
behaviors in rats, electrical stimulation
augments these behaviors, and androgen
implants in this region in castrates reinstate the behaviors. Significantly, cell bodies of neurons in the MPOA concentrate
radioactivity heavily after injections of tritiated estradiol, testosterone, or dihydrotestosterone in autoradiographic studies
(e.g., Pfaff, 1968; Stumpf, 1968; Stumpf
and Sar, 1978).
The evidence for androgen action in the
lower spinal cord to modulate sexual reflexes is nearly as strong as the evidence
for its action in the MPOA. Hart (1978a,
b) has demonstrated that male rats possess
penile flip reflexes which are elicited by
holding back the preputial sheath of the
penis. These flips are decreased by castration and enhanced by administration of
testosterone propionate, and they occur
more frequently following mid-thoracic
transection of the spinal cord. Implants of
testosterone propionate in the spinal cord
increase the flips in castrated males (Hart
and Haugen, 1969), suggesting that testosterone propionate or its metabolites act in
the spinal cord to modulate these reflexes.
It is unclear to what extent these reflexes,
elicited by artificial stimulation, are similar
to reflexes during normal copulatory behavior. However, even if the reflexes are
labeled "artificial," it is quite likely that
they are controlled by neural circuits
which are normally involved in erection,
intromission, and/or ejaculation. Sachs and
Garinello (1978) and Kurtz and Santos
(1979) have demonstrated that males exposed to a female and allowed to initiate
a series of intromissions will show much
shorter latencies for penile reflexes than
controls not exposed to females, and that
sexual exhaustion depresses all such reflexes. These studies support the notion
that the penile reflexes are a measure of
the function of neural circuits involved in
copulation, and give credence to the assertion that androgens act on the spinal conj
to modulate copulatory behavior.
To search for cellular accumulation of
androgen which would correlate with a
spinal action of androgen on this system,
Breedlove and Arnold (1980) located the
motor neurons which innervate the striated muscles of the penis, which are almost
certainly those responsible for the penile
flip reflexes. When horseradish peroxidase
is injected into the bulbocavernosus or levator ani muscles, it is retrogradely transported to a discrete motor nucleus in the
ventromedial lumbar spinal cord. In autoradiograms, these motor neurons accumulate hormone after injection of tritiated
testosterone, dihydrotestosterone, but not
after estradiol (Breedlove and Arnold,
1980, which extends earlier work of Sar
and Stumpf, 1977). This pattern of accumulation agrees perfectly with the pattern
of steroid activation of the penile reflexes,
since of these three hormones, only estradiol is ineffective in reinstating the reflexes
in castrated males (Hart, 1978a). These results raise the likelihood that androgens
act on these motor neurons to influence
the function of neural circuits involved in
some aspect of the genital reflexes demonstrated by Hart, and hence in copulatory behavior. The autoradiographic studies located androgen-labeled cells in other
spinal nuclei, including lamina X, intermediolateral nucleus, and the ventrolateral motor neurons, among which are
found the motor neurons of the ischiocavernosus. Thus there are several potential
sites of action of androgens even within
the spinal cord.
At present we do not know precisely
how the penile striated muscles (bulbocavernosus, ischiocavernosus, and levator ani)
participate in male copulatory behavior in
rats. However, in dogs the ischiocavernosus and bulbospongiosus (the homolog of
the bulbocavernosus) are responsible for
vascular pressure increases in the penis
which mediate erection and copulatory
LEVELS OF STEROID ACTION ON BEHAVIOR
lock in that species (Purohit and Beckett,
1976). In the stallion, these muscles contract during intromission and ejaculation
(Beckett et al., 1975). Thus it is likely that
these muscles alter erectile mechanisms in
0ats, too. However, in the human male the
homologs of these muscles are also involved in control of micturation (Warwick
and Williams, 1973), so that we cannot be
certain that the motor neurons that accumulate androgen are activated exclusively
during copulation in rats.
In addition to effects of androgen on the
CNS, Beach and his co-workers have suggested a peripheral action on the penis
which may be important in copulatory behavior (Beach and Levinson, 1950; Beach,
1971). This is based primarily on a strong
correlation between testosterone's influence on the amount of copulatory behavior and its influence on the size and morphology of the cornified papillae on the
glans penis, which are thought to be important for its tactile sensitivity. It is known
that gross alterations in tactile sensitivity of
the penis produced by topical anesthetics
will lead to decrements in ability to intromit (reviewed by Malsbury and Pfaff,
1974), but there seems to be no direct evidence that the sensitivity of the glans penis
is altered by androgens. Indeed, Cooper
and Aronson (1974) could detect no sensory effect of androgens on the penis of
cats. It is possible to administer non-aromatizable androgens at moderate doses2
(fluoxymesterone or dihydrotestosterone)
to castrates, which affect the morphology
of the penile papillae but have little effect
on copulatory behavior or penile reflexes
(Beach and Westbrook, 1968; Feder, 1971;
Hart, 1978a). While these results indicate
that androgenic effects on the papillae are
not sufficient to reinstate copulatory behavior or penile reflexes, they have little
bearing on the question of whether androgens modulate the sensitivity of the penis.
Although most authors interested in an2
Dihydrotestosterone does have effects on penile
reflexes when given in higher doses (Hart, 1978a,
1979) and on copulatory behavior when given in
combination with estradiol (Davis and Barfield,
1979).
237
drogen-induced copulatory behavior have
considered changes in the sensitivity of the
penis in detail, few have considered direct
androgenic effects on the penile striated
muscles as possibly important for copulatory behavior. These muscles are among
the most androgen-sensitive muscles in the
body of the rat. Castration reduces the sarcoplasmic mass of the levator ani (Venable,
1966), and castration and administration
of androgen cause a host of biochemical
and physiological changes in this muscle
which alter its size and functional capacity
(Buresova and Gutmann, 1971; Tucek et
al., 1976; Vyskocil and Gutmann, 1977;
Hanzlikova and Gutmann, 1978). Although some of these steroid effects could
be exerted via an action on the androgensensitive motor neurons of these muscles
(Buresova et al., 1972), at least some of the
metabolic and functional effects have been
demonstrated in cultured muscles, suggesting direct actions on the muscle (Buresova and Gutmann, 1971). The levator
ani and bulbocavernosus appear to be virtually identical in their biochemical affinity
for androgens (Krieg et al., 1974). Each
possesses high affinity receptors for testosterone and dihydrotestosterone, and the
numbers of these receptors is 5-7 times
greater than in other striated muscles
(Tremblay et al., 1977; Krieg and Voight,
1977). Both dihydrotestosterone and
testosterone are effective in stimulating
growth of the levator ani (Liao and Fang,
1969). Thus, these muscles could be one
site of action of androgens in their influences on male copulatory behavior.
If one were to list the sites of action of
androgen related to copulatory behavior
in the order of our current certainty regarding their involvement, the list might
appear as follows: (1) MPOA, (2) spinal
cord, possibly at the level of the motor
neurons controlling penile striated muscles, (3) the penile striated muscles themselves, and (4) certain steroid-concentrating limbic brain regions. The evidence is
clearest for the MPOA. It is weaker for the
penile motor neurons and their target
muscles because we have little idea so far
how implants of steroids in these places
would affect copulatory behavior. Yet
238
ARTHUR P. ARNOLD
N
SYRIN
FIG. 1. Some of the brain regions thought to be involved in vocal control in zebra finches are shown in
highly schematic form in this parasagittal view, together with anatomical projections discovered in the
canary by Nottebohm et al. (1976) and Nottebohm
and Kelley (1978). Black dots indicate the presence
of cells labeled by androgens (Arnold et al., 1976;
Arnold, 1979, 19806).
there can be little doubt that these structures play an important role in copulation.
The evidence for copulation-related effects of androgen on other CNS areas is
rather speculative at present.
ANDROGEN EFFECTS ON BIRD SONG
The second behavioral system to be discussed is the neural system involved in the
control of song and other vocalizations in
the zebra finch. Zebra finch males sing a
quiet courtship song which is oriented toward the female, and together with specific patterns of movement, body orientation,
and feather erection, serves as an attractive
stimulus for females. Song is reduced in its
occurrence after castration, and androgens reinstate it (Prove, 1974; Arnold,
1975). This behavior is controlled by an
interconnected system of discrete neural
regions (Fig. 1) which were originally described in the canary (Nottebohm et al.,
1976; Nottebohm and Kelley, 1978; Kelley
and Nottebohm, 1979). Lesions of the hyperstriatum ventrale pars caudale (HVc)
and nucleus robustus archistriatalis (RA)
disrupt song in the canary. HVc neurons
project to RA, and RA in turn projects to
nucleus intercollicularis (ICo), which has
long been implicated in vocal control in
many bird species (e.g., Phillips and Peek,
1975), and to the tracheosyringeal motor
neurons of the hypoglossus (nXIIts),
which innervate the muscular vocal organ,
or syrinx. Two other brain regions, the
magnocellular nucleus of the anterior
neostriatum (MAN) and Area X of the 1^
bus parolfactorius are also thought to be
involved in song control since they have
heavy monosynaptic connections with
either HVc or RA (Fig. 1). The connections of the song system described in the
canary by Nottebohm and his colleagues
have been confirmed in the zebra finch (M.
E. Gurney, personal communication; Lewis et al., 1979; Arnold, 1980a).
In suggesting where androgens might
act to influence singing behavior in zebra
finches, we must at present rely almost exclusively on autoradiographic data which
establish which brain regions involved in
song contain cells which specifically accumulate sex steroids. Of the brain regions
shown in Figure 1, HVc, MAN, RA, and
nXIIts contain neurons which accumulate
hormone after injection of tritiated testosterone or dihydrotestosterone, but not after estradiol (Arnold et al., 1976; Arnold,
1979, 19806), and ICo contains labeled
cells after injection of any of these. Furthermore, the syringeal muscles themselves are quite sensitive to androgens.
Castration reduces the size of these muscles, and testosterone propionate administration reverses this effect (Arnold, 1974;
Luine et al., 1980). The syringeal tissues
contain specific high affinity androgen
binding proteins (Lieberburg and Nottebohm, 1979). The activities of acetylcholinesterase and choline acetyltransferase in
the syrinx and tracheosyringeal nerve are
regulated by androgen titers in the blood
(Bleisch et al., 1979). Although it is not
clear at present precisely how much of the
androgen effects on these enzyme activities in the syringeal muscles can be attributed to an effect on the motor neurons of
nXIIts which innervate the muscles, it is
quite likely that there is a direct action on
the muscles themselves, in view of the high
levels of androgen receptors in the syrinx
itself.
When considering this evidence in the
LEVELS OF STEROID ACTION ON BEHAVIOR
aggregate it is tempting to suggest that as
in the neural systems controlling the ovulatory surge of LH and male rat copulatory
behaviors, (a) sex steroids may act at many
levels in the song system to influence this
^jehavior, and (b) one of these levels is the
™
final common pathway of the system, i.e.,
the motor neurons controlling song and
the syringeal muscles themselves. However, we must hesitate before accepting
this conclusion fully at present, since for
all of the neural regions which accumulate
androgens except nXIIts, we have no direct evidence that androgens affect the
function of these neurons, and in no case
is it apparent that such an action actually
is important for vocal behavior. Nevertheless, the parallels between these three
neural systems are sufficiently great that
this lends weight to the notion that multilevel actions of steroids may not be uncommon in the control of neuroendocrine and
behavioral functions.
In fact, Pfaff and Modianos (1980) have
suggested that estrogen exerts its effects
on lordosis behavior of the female rat via
actions on neurons in various interconnected regions of the brain, each of which
is influential in control of lordosis. Similarly, Kelley (1980) has presented autoradiographic evidence of multi-level sites of
action of androgens in circuits controlling
reproductive vocalizations of the frog Xenopus laevis. In these two neural systems,
there is evidence that sex steroids act on
structures which carry relevant sensory information. This is clearest in the rat, where
estradiol increases the size of receptive
fields of sensory neurons innervating the
perineal region of the female (Komisaruk
et al, 1972; Kow and Pfaff, 1973). Since
this influence of estradiol is detected even
when the pudendal nerve carrying the sensory fibers is severed from the CNS, this
effect does not depend on centrifugal influences of the CNS on the periphery. In
Xenopus, a mesencephalic auditory nucleus
contains cells that accumulate sex steroids,
and this accumulation could mediate possible androgen influences on auditory processes relevant to Xenopus mating calls
)
(Kelley, 1980). Significantly, other brain
239
regions implicated in the control of calling
accumulate steroids, including the laryngeal motor neurons (Kelley et al., 1975;
Kelley, 1980).
If sex steroids do act at many levels in
circuits controlling behavior this might
seem to imply that there is considerable
redundancy in hormonal activation of a
particular behavior. To support this view,
it is possible to point to studies which demonstrate that implants of testosterone into
the MPOA of male rats activate male sex
behavior. If an implant at one locus in the
neural circuit is sufficient to activate the
behavior in a castrate, then it would seem
that the other sites could only serve a redundant function. This view may be incorrect. Because there may be leakage of
the hormone to sites distant from the implants, even in studies which look for and
fail to find such leakage, the activation of
the behavior may depend on marginal steroid effects on such distant sites. Furthermore, because copulatory behavior evoked
by such implants is rarely normal in its topography (Hart, 1974), other sites of action can be invoked. It is quite conceivable
that at each level, the steroid exerts a specific effect which is not functionally equivalent to that exerted at any other level.
This should not be confused with redundancy.
To turn more specifically to the question
posed at the beginning of this paper, what
can be said about the logical levels at which
steroids exert their action? At present we
can have only the barest glimpse of an answer to this intriguing question: Since the
anatomical sites of steroid action are likely
to be multiple and diverse, so are the logical functions performed by these sites.
The only sites of action which have a clear
logical function in the three systems discussed here are the motor neurons and
muscles in the behavioral systems, and the
pituitary in the LH control system. Each of
these is part of the final common path.
Yet the motor neuronal and muscular
action of steroids were what Tinbergen
(1951) specifically thought would be unlikely. As discussed above, there seem to
be two main considerations which led Tin-
240
ARTHUR P. ARNOLD
bergen, quite logically, to the speculation
that steroids must modify the function
only of centers which are high in the hierarchy. (1) The action of hormones on
motor neurons and muscles seems to be
incompatible with a specific action on a
particular behavior, since any given muscle
is often involved in more than one behavior. (2) It seems unlikely that the "switching" function of hormones (i.e., increasing
the frequency of certain behaviors at the
expense of others) would be exerted at the
motor neuronal or muscular level. For example, if this were the only site of action
in the zebra finch song system, and androgens modified the frequency of the behavior at the motor neuronal level, one is
faced with the improbable notion of a castrate male "singing to himself." That is, the
neuronal pattern for song would be generated in the brain but is not expressed in
muscular contractions of the syrinx because of a lack of androgen effects on the
motor neurons and muscle fibers.
It is possible to reconcile these two
points with our present knowledge of sites
of steroid action. With regard to the first
point, it is true that the syringeal muscles
of zebra finches are active during both
song control and breathing, and in rats the
penile muscles could be involved in the
control of micturation, in addition to their
sexual function. Yet it would still be possible for androgens to act on the syringeal
motor neurons and muscles to influence
singing without also markedly changing
the function of the muscles during breathing. For example, if androgens increase
the activity of choline acetyltransferase in
the hypoglossal motor neurons (Bleisch et
al, 1979), this might allow them to synthesize enough acetylcholine to keep pace
with increased use of these neurons which
requires more discharge of acetylcholine.
Yet the amount of acetylcholine release
per impulse in the motor neurons might
stay constant so that there would be no effect on the ability of these neurons to function properly when driving the muscles
during breathing. Although this is highly
speculative, it serves to demonstrate that
certain effects of androgens at this level
need not be non-specific. It also helps to
resolve the second consideration, since one
can imagine androgen effects on motor
neurons which would not play a role in
altering the frequency of a behavior (i.e.,
a switching role), but would merely alloiQ
the motor neurons and muscles to handle
the increase in usage demanded if there is
a switch in the behavioral repertoire.
Thus, Tinbergen was probably right,
that hormones are most likely to switch
behaviors on or off by acting centrally, at
logically high levels in the neural circuitry.
Yet there may be other, non-switching
functions played by steroids, which can occur at any level. There is currently great
need for detailed studies of the changes
exerted by steroids at each level, so that a
better picture may emerge of the logical
functions which are modulated by steroids.
ACKNOWLEDGMENTS
The author's research has been supported by NSF grant BNS 7705973 and
USPHS grant 5-SO7 RR07009-14.
REFERENCES
Arnold, A. P. 1974. Behavioral effects of androgen
in zebra finches (Poephila gutlata) and a search
for its sites of action. Ph.D. Diss., The Rockefeller
University.
Arnold, A. P. 1975. The effect of castration and androgen replacement on song, courtship, and
aggression in zebra finches (Poephila guttala). J.
Exp. Zool. 191:309-326.
Arnold, A. P. 1979. Hormone accumulation in the
brain of the zebra finch after injection of various
steroids and steroid competitors. Soc. Neurosci.
Abstr. 5:434.
Arnold, A. P. 1980a. Sexual differences in the brain.
Amer. Sci. 68:165-172.
Arnold, A. P. 19804. Quantitative analysis of sex differences in hormone accumulation in the zebra
finch brain: Methodological and theoretical issues. J. Comp. Neurol. 189:421-436.
Beach, F. A. 1971. Hormonal factors controlling the
differentiation, development, and display of copulatory behavior in the ramstergig and related
species. In E. Toback, L. R. Aronson, and E.
Shaw (eds.), Biopsychology of development, pp. 2 4 9 -
296. Academic Press, New York.
Beach, F. A. and G. Levinson. 1950. Effects of androgen on the glans penis and mating behavior
of castrated male rats. J. Exp. Zool. 114:159-171.
Beach, F. A. and W. H. Westbrook. 1968. Dissociation of androgenic effects on sexual morphology
and behavior in male rats. Endocrinology
83:395-398.
LEVELS OF STEROID ACTION ON BEHAVIOR
Beck, L. V., M. Bay, A. F. Smith, D. King, and R.
Long. 1978. Steroid priming of the luteinizing
hormone response to luteinizing hormone releasing hormone. J. Endocrinol. 77:293-299.
Beckett, S. D., D. F. Walker, R. S. Hudson, T. M.
Reynolds, and R. C. Purohit. 1975. Corpus
A
spongiosum penis pressure and penile muscle
activity in the stallion during coitus. Am. J. Vet.
Res. 36:431-433.
Blake, C. A. 1978. Neurohumoral control of cyclic
pituitary LH and FSH release. Clinics in Obstetrics and Gynaecology 5:305-327.
Bleisch, W., V. Luine, F. Nottebohm, and B. McEwen. 1979. Testosterone effects on choline
acetylase and acetylcholinesterase in the Xllth
cranial nerve and syringeal muscle of the zebra
finch. Soc. Neurosci. Abstr. 5:476.
Breedlove, S. M. and A. P. Arnold. 1980. Hormone
accumulation in a sexually dimorphic motor nucleus of the rat spinal cord. Science 210:564566.
Brown-Grant, K. 1977. Physiological aspects of the
steroid hormone-gonadotropin interrelationship. In R. O. Greep (ed.), Reproductive physiology
II, Int. Rev. Physiol. 13:57-83. University Park
Press, Baltimore.
Buresova, M. and E. Gutmann. 1971. Effect of testosterone on protein synthesis and contractility
of the levator ani muscle of the rat. J. Endocrinol.
50:643-651.
Buresova, N., E. Gutmann, and V. Hanzlikova. 1972.
Differential effects of castration and denervation
on protein synthesis in the levator ani muscle of
the rat. J. Endocrinol. 54:3-14.
Chiappa, S. A., G. Fink, and N. M. Sherwood. 1977.
Immunoreactive luteinizing hormone releasing
factor (LRF) in pituitary stalk plasma from female rats: Effects of stimulating diencephalon,
hippocampus, and amygdala. J. Physiol. 267:625640.
Cooper, K. K. and L. R. Aronson. 1974. Effects of
castration on neural afferent responses from the
penis of the domestic cat. Physiol. and Behav.
12:93-107.
Cramer, O. N. and C. A. Barraclough. 1971. Effect
of electrical stimulation of the preoptic area on
plasma LH concentrations in proestrus rats. Endocrinology 88:1175-1183.
Davis, P. G. and R. J. Barfield. 1979. Activation of
masculine sexual behavior by intracranial estradiol benzoate implants in male rats. Neuroendocrinology 28:217-227.
Feder, H. A. 1971. The comparative actions of testosterone propionate and 5 alpha-androstan-17
beta-ol-3 one propionate on the reproductive behavior, physiology and morphology of male rats.
J. Endocrinol. 51:241-252.
Fink, G. 1979. Feedback actions of target hormones
on hypothalamus and pituitary with special reference to gonadal steroids. Ann. Rev. Physiol.
41:571-585.
Goodman, R. L. 1978. The site of the positive feedback action of estradiol in the rat. Endocrinology 102:151-159.
241
Hanzlikova, V. and E. Gutmann. 1978. Effect of castration and testosterone administration on the
neuromuscular junction in the levator ani muscle
of the rat. Cell Tiss. Res. 189:155-166.
Hart, B. L. 1974. Gonadal androgen and sociosexual
behavior of male mammals: A comparative analysis. Psych. Bull. 81:383-400.
Hart, B. L. 1978a. Reflexive mechanisms in copulatory behavior. In T. E. McGill, D. A. Dewsbury,
and B. D. Sachs (eds.), Sex and behavior, pp. 205242, Plenum, New York.
Hart, B. L. 1978*. Hormones, spinal reflexes, and
sexual behavior. In J. B. Hutchison (ed.), Biological determinants of sexual behavior, pp. 319-348.
Wiley, New York.
Hart, B. L. 1979. Activation of sexual reflexes of
male rats by dihydrotestosterone but not estrogen. Physiol. and Behav. 23:107-109.
Hart, B. L. and C. M. Haugen. 1968. Activation of
sexual reflexes in male rats by spinal implantation of testosterone. Physiol. and Behav. 3:735738.
Kawakami, M., E. Terasawa, F. Kimura, and K. Wakabayashi. 1973. Modulatory effect of limbic
structures on gonadotropin release. Neuroendocrinol. 12:1-16.
Kelley, D. B. 1980. Auditory and vocal nuclei of frog
brain concentrate sex hormone. Science 207:553555.
Kelley, D. B., J. I. Morrell, and D. W. Pfaff. 1975.
Autoradiographic localization of hormone-concentrating cells in the brain of an amphibian
Xenopus laevis. I. Testosterone. J. Comp. Neurol. 164:47-61.
Kelley, D. B. and F. Nottebohm. 1979. Projections
of a telencephalic auditory nucleus—Field L—in
the canary. J. Comp. Neurol. 183:455-470.
Komisaruk, B. R., N. T. Adler, and J. B. Hutchison.
1972. Genital sensory field: Enlargement by estrogen treatment in female rats. Science 178:12951298.
Kow, L.-M. and D. W. Pfaff. 1973. Effects of estrogen treatment on the size of receptive field and
response threshold of pudendal nerve in the female rat. Neuroendocrinology 13:299-313.
Krieg, M., R. Szalay, and K. D. Voigt. 1974. Binding
and metabolism of testosterone and of 5 alphadihydrotestosterone in bulbocavernosus/levator
ani (BCLA) of male rats: In vivo and in vitro
studies. J. Steroid Biochem. 5:453^159.
Krieg, M. and K. D. Voigt. 1977. Biochemical substrate of androgenic actions at a cellular level in
prostate, bulbocavernosus/levator ani and in
skeletal muscle. Acta Endocrinol. Suppl. 85:4389.
Kurtz, R. G. and R. Santos. 1979. Supraspinal influences on the penile reflexes of the male rat: a
comparison of the effects of copulation, spinal
transection, and cortical spreading depression.
Horm. Behav. 12:73-94.
Larsson, K. 1979. Features of the neuroendocrine
regulation of masculine sexual behavior. In C.
Beyer (ed.), Endocrine control of sexual behavior,
pp. 77-163. Raven Press, N.Y.
242
ARTHUR P. ARNOLD
Lewis, J. W., S. M. Ryan, L. L. Butcher, and A. P.
Arnold. 1979. Identification of catecholaminergic projections to Area X in the zebra finch.
Soc. Neurosci. Abstr. 5:143.
Liao, S. and S. Fang. 1969. Receptor-protein for androgens and the mode of action of androgens on
gene transcription in ventral prostate. Vitam.
Horm. 27:17-90.
Lieberburg, I. and F. Nottebohm. 1979. High-affinity androgen binding proteins in syringeal tissues
of song birds. Gen. Comp. Endocrinol. 37:286293.
Luine, V., F. Nottebohm, C. Harding, and B. S. McEwen. 1980. Androgen affects cholinergic enzymes in syringeal motor neurons and muscle.
Brain Res. 192:89-107.
Malsbury, C. W. and D. W. Pfaff. 1974. Neural and
hormonal determinants of mating behavior in
adult male rats. A review. In L. D. DiCara (ed.),
Purohit, R. C. and S. D. Beckett. 1976. Penile pressures and muscle activity associated with erection
and ejaculation in the dog. Am. J. Physiol.
231:1343-1348.
Sachs, B. D. 1978. Conceptual and neural mechanisms of masculine copulatory behavior. In T. E.
McGill, D. A. Dewsbury, and B. D. Sachs (eds.)^
Sex and behavior, pp. 267—296. Plenum Press,
N.Y.
Sachs, B. D. and L. D. Garinello. 1978. Interaction
between penile reflexes and copulation in male
rats. J. Comp. Physiol. 92:759-767.
Sar, M. and W. E. Stumpf. 1977. Androgen concentration in motor neurons of cranial nerves and
spinal cord. Science 197:77—79.
Sarkar, D. K., S. A. Chiappa, and G. Fink. 1976.
Gonadotropin-releasing hormone surge in proestrus rats. Nature 264:461-462.
Schwartz, N. B. 1972. Reproduction: Gonadal funcLimbic and autonomic nervous system research, pp.
tion and its regulation. Ann. Rev. Physiol.
85-136. Plenum Press, N.Y.
34:425-472.
McF.wen, B. S. 1976. Gonadal steroid receptors in Stumpf, W. E. 1968. Estradiol concentrating neuneuroendocrine tissues. In F. Naftolin et al.
rons: Topography in the hypothalamus by drymount autoradiography. Science 162:1001-1003.
(eds.), Subcellular mechanisms in reproductive neuroendocrinology, pp. 277-304. Elsevier, Amster- Stumpf, W. E. and M. Sar. 1978. Anatomical distridam.
bution of estrogen, androgen, progestin, corticosteroid, and thyroid hormone target sites in
Nottebohm, F. and D. B. Kelley. 1978. Projections
the brain of mammals: Phylogeny and ontogeny.
to efferent vocal control nuclei of the canary telAmer. Zool. 18:435-445.
encephalon. Soc. Neurosci. Abstr. 4:101.
Nottebohm, F., T. M. Stokes, and C. M. Leonard. Tinbergen, N. 1951. The study of instinct. Oxford
1976. Central control of song in the canary, SeUniv. Press, N.Y.
rinus canarius. J. Comp. Neurol. 165:457-486.
Tremblay, R. R., J. Y. Dube, M. A. Ho-Kim, and R.
Lesage. 1977. Determination of rat muscle anPfaff, D. W. 1968. Autoradiographic localization of
drogen-receptor complexes with methyltrienoradioactivity in rat brain after injection of tritiated sex hormones. Science 161:1355-1356.
lone. Steroids 29:185-195.
Pfaff, D. W. and M. Keiner. 1973. Atlas of estradiol Tucek, S., D. Kostirova, and E. Gutmann. 1976. Tesconcentrating cells in the central nervous system
tosterone induced changes of choline acetylof the female iat.j. Comp. Neurol. 151:121-158.
transferase and cholinesterase activities in rat levator ani. J. Neurol. Sci. 27:353-362.
Pfaff, D. W. and D. Modianos. 1980. Neural mechanisms of female reproductive behavior. In R. Velasco, M. E. and S. Taleisnik. 1969. Effect of hipGoy and D. Pfaff (eds.), Neurobiology of reproducpocampal stimulation on the release of gonadotion. Plenum Press, N.Y. (In press)
tropin. Endocrinology 85:1154-1159.
Phillips, R. E. and F. W. Peek. 1975. Brain organi- Venable,J. H. 1966. Morphology of the cells of norzation and neuromuscular control of vocalization
mal, testosterone deprived and testosterone stimin birds. In P. Wright, P. G. Caryl, and D. M.
ulated levator ani muscles. Am. J. Anat.
Vowles (eds.), Neural and endocrine aspects of be119:271-302.
havior in birds, pp. 243-274. Elsevier, Amster- Vyskocil, F. and E. Gutmann. 1977. Electrophysiodam.
logical and contractile properties of the levator
Prove, E. 1974. Der Einfluss von Kastration und
ani muscle after castration and testosterone
Testosteronsubstitution auf das Sexualverhalten
administration. Pfluegers Arch. 368:105-109.
mannlicher Zebrafinken (Taeniopygia guttata cas- Warwick, R. and P. L. Williams. 1973. Gray's anatomy.
tanotis Gould). J. fuer Ornithologie 115:338-347.
35th British edition, p. 531. W. B. Saunders,
Philadelphia.