The role of sex steroids in the acquisition and production of birdsong

Endocrine modulation of the development and adult function of the avian song system

Enough data are now available on the neurobiology of the avian song system and on the development and performance of song that sophisticated questions on the relations between the behavior and the neurobiology can be addressed. This review describes what is known of sex differences and individual differences in the neurobiology of the song system in mature birds. It summarizes data on the role of steroid hormones in the development of the song system and what is known of steroid-related adult plasticity in this system. Finally, it discusses hypotheses on the relations between structure and function in this system and suggests issues that must be addressed in future studies.

Testosterone metabolism in the avian hypothalamus

Many central actions of testosterone (T) require the transformation of T into several metabolites including 5 alpha-dihydrotestosterone (5 alpha-DHT) and estradiol (E2). In birds as in mammals, 5 alpha-DHT and E2, alone or in combination, mimic most behavioral effects of T. The avian brain is, in addition, able to transform T into 5 beta-DHT, a metabolite which seems to be devoid of any behavioral or physiological effects, at least in the context of reproduction. By in vitro product-formation assays, we have analyzed the distribution, sex differences and regulation by steroids of the 3 main T metabolizing enzymes (aromatase, 5 alpha- and 5 beta-reductases) in the brain of the Japanese quail (Coturnix c. japonica) and the zebra finch (Taeniopygia guttata castanotis). In the hypothalamus of quail and finches, aromatase activity is higher in males than in females. It is also decreased by castration and increased by T. The activity of the 5 alpha-reductase is not sexually differentiated nor controlled by T. The 5 beta-reductase activity is often higher in females than in males but this difference disappears in gonadectomized birds and no clear effect of T can be observed at this level. The zebra finch brain also contains a number of steroid-sensitive telencephalic nuclei [e.g. hyperstriatum ventrale, pars caudale (HVc) and robustus archistriatalis (RA)] which play a key role in the control of vocalizations. These nuclei also contain T-metabolizing enzymes but the regulation of their activity is substantially different from what has been observed in the hypothalamus. Aromatase activity is for example higher in females than in males in HVc and RA and the enzyme in these nuclei is not affected by castration nor T treatment. In these nuclei, the 5 alpha-reductase activity is higher in males than in females and the reverse is true for the 5 beta-reductase. These sex differences in activity are not sensitive to gonadectomy and T treatment and might therefore be organized by neonatal steroids. We have been recently able to localize aromatase-immunoreactive (AR-ir) neurons by ICC in the brain of the quail and zebra finch. Positive cells are found in the preoptic area, ventromedial and tuberal hypothalamus. AR-ir material is found in the perikarya of cells and fills the entire cellular processes including axons. At the electron microscope level, immunoreactive material can clearly be observed in the synaptic boutons. This observation raises questions concerning the mode of action of estrogens produced by central aromatization of T.

Song acquisition in photosensitive and photorefractory male European starlings

We tested the ability of 1-year-old European starlings (Sturnus vulgaris) to acquire songs while in different physiological states. Photorefractory males, with low testosterone levels, learned songs as completely and as accurately as photosensitive males in full reproductive condition. This indicates that song acquisition in 1-year-old males does not depend on high levels of androgens. The ability to learn songs during the nonbreeding season may reflect the high song output of potential tutors through most of the year, including the photorefractory period, and may facilitate increases in song repertoire size in adulthood.

Song learning: the interface between behaviour and neuroethology

The high degree of developmental plasticity displayed by the songs of oscine birds makes them appropriate subjects for research on the etiology and neurobiology of vocal learning. Strong individual differences and learned local dialects are common. The readiness to acquire new songs appears to persist throughout life in some species and is restricted to relatively short sensitive periods in others. Learning can occur with remarkably few exposures to song. Mimicry of other species occurs but, given a choice, there is a tendency to favour conspecific songs. Evidence is presented for two kinds of vocal learning, one 'memory-based', the other 'action-based.' Subsong and 'plastic song' phases of motor development appear to be obligatory steps in the ontogeny of learned songs. A case is made that acquisition and production should be viewed as distinct phenomena with different physiological correlates. Research on behavioural development is closely associated with studies of the physiology of development. The two are mutually synergistic, and the synergism is well displayed in research on song learning in birds. This review of some of the characteristics of avian vocal learning as derived from behavioural studies, indicates lacunae in our knowledge about the ethology of song learning, and suggests how the comparative study of vocal development can pave the way for new insights into the underlying neurobiology.

Human chorionic gonadotropin-induced shifts in the electrosensory system of the weakly electric fish, Sternopygus

Sternopygus macrurus of both sexes were injected with human chorionic gonadotropin (hCG) or saline. Electric organ discharge (EOD) frequency rose after hCG injections in females and gradually declined to baseline levels over the next few weeks. EOD changes in males were more complex and variable; most males showed an initial minor rise in EOD frequency followed by a larger decrease, or simply a decrease. hCG treatment also resulted in a rise in electroreceptor best frequency and shortened electric organ pulse duration in females, and had the opposite effect on these parameters in males. The saline-injected controls showed no changes in any of these parameters. Levels of testosterone (T) and 11-ketotestosterone, but not estrogen (E), were elevated in males preceding the fall in EOD frequency, whereas neither T nor E changed significantly in females before EOD frequency increases. Saline injections caused a drop in T in the male control group and had no effect in the female control group. We presume that the effect of hCG on the electrosensory system of males is mediated via androgens. Whether the effects of hCG on females are mediated by slight increases in circulating levels of gonadal steroids, the release of hormones other than T or E, or are due to direct effects on the nervous system is not known.

Effects of castration and testosterone treatment on the activity of testosterone-metabolizing enzymes in the brain of male and female zebra finches

Recently, we described the distribution of testosterone-metabolizing enzymes (i.e., aromatase, 5 alpha- and 5 beta-reductases) in the zebra finch (Taeniopygia guttata) brain using a sensitive radioenzyme assay combined to the Palkovits punch method. A number of sex-differences in the activity of these enzymes were observed especially in nuclei of the song-control system. The hormonal controls of these differences have now been analyzed by gonadectomizing birds of both sexes and by giving them a replacement therapy with silastic implants of testosterone (T). Five nuclei of the song system (Area X [X], nucleus magnocellularis of the anterior neostriatum [MAN], nucleus robustus archistriatalis [RA], nucleus intercollicularis [ICo], hyperstriatum ventrale, pars caudalis [HVc]) and three preoptic-hypothalamic areas (preoptic anterior [POA], periventricular magnocellular nucleus [PVM], and posterior medial hypothalamic nucleus [PMH]) were studied as well as other limbic and control non-steroid-sensitive areas. The activity of the 5 alpha-reductase was higher in males than in females for the five song-control nuclei and was not affected by the hormonal treatments. The overall activity of this enzyme was not sexually dimorphic in POA and PVM. It was higher in males than in females in intact birds only, and was reduced by gonadectomy and enhanced by T. The activity of the 5 beta-reductase was higher in females than in males in all nuclei of the song system and in POA, but was not influenced by the changes in T level. Both sex and treatment effects were observed in the control of aromatase. The production of estrogens was dimorphic (females greater than males) in RA and PMH. It was increased by T in POA, PVM, and PMH, and also in RA. These data show that some of the sex differences in T-metabolizing enzymes result from the exposure to different levels of T in adulthood (e.g., 5 alpha-reductase in POA and PVM or aromatase in PVM), whereas others persist even if birds are exposed to the same hormonal conditions. These are presumably the result of organizational effects of steroids. The steroid modulation of the aromatase might be related directly to the activation of sexual, aggressive, and nest-building behaviors, whereas the stable dimorphism in 5 alpha- and 5 beta-reductase observed in the nuclei of the song system might be one of the neurochemical bases of the sex differences in the vocal behavior of the zebra finch.

Sex steroid levels in developing and adult male and female zebra finches (Poephila guttata)

Serum samples from male and female zebra finches ranging in age from 1 day before hatch to 54 days posthatch were assayed for 17 beta-estradiol (E), androgen, testosterone (T), or 5 alpha-dihydrotestosterone (DHT). Additional samples were assayed from intact and gonadectomized adults, gonadectomized adults with intraperitoneal implants of testosterone propionate (TP) or estradiol benzoate (EB), gonadectomized nestlings, and nestlings injected subcutaneously with EB. DHT levels of developing birds did not vary as a function of either sex or age. During development, average androgen and T levels were highest during the nestling period, prior to sexual maturation, and were higher in females than in males. Endogenous androgen levels of most subjects that were sampled repeatedly rose and then declined between 24 and 49 days. TP implants produced higher T levels in adult females than in adult males. Levels of E were higher in both sexes during the hatching period (Days -1 through 0) than during the nestling period (Days 2 through 14). A greater number of males than females had relatively high E levels on Days 12 and 14 and during the second week overall. There was no sex difference in levels of E in adults, and gonadectomized adults had markedly higher E levels than intact adults. Gonadectomized nestlings had the same androgen and E levels as intact nestlings of the same age; EB injected nestlings had elevated E levels. The present results indicate specific endocrine changes that mirror events crucial to sexual differentiation of endocrine and behavioral components of reproduction, and have important implications for models of sexual differentiation in zebra finches.

Sexual differentiation of behavior in the zebra finch: effect of early gonadectomy or androgen treatment

Treatment of nestling zebra finches with estradiol benzoate (EB) has been shown to masculinize singing in females and demasculinize copulatory behavior in males, suggesting that sexual differentiation of these behaviors is under hormonal control such that testicular hormones induce the capacity for song and ovarian hormones suppress the capacity for mounting. Two experiments were carried out to obtain a more complete picture of sexual differentiation in this species. In Experiment 1, nestlings were injected daily for the first 2 weeks after hatching with testosterone propionate (TP), dihydrotestosterone propionate (DHTP), or a combination of DHTP and EB. As adults, birds were gonadectomized and implanted with TP prior to testing, then tested again after implantation with EB. Singing was not increased in females by any of the treatments. The only effect of either TP or DHTP given alone was defeminization of female proceptive behavior by DHTP. Thus androgens appear to have less influence than estrogens on sexual differentiation of behavior in this species. The combination of DHTP and EB demasculinized mounting in males. In Experiment 2, nestlings were gonadectomized at 7-9 days of age and implanted with TP prior to testing in adulthood. Early gonadectomy had little effect on later behavior; early castrated males sang, danced, and copulated normally and early ovariectomized females neither sang nor mounted.

Neurophysiological and behavioral development in birds: song learning as a model system

The avian song system is a particularly good model for studying the behavioral and physiological aspects of animal development. One seemingly trivial but very important reason for this is that the sound spectrograph enables sounds to be described, measured and analyzed objectively and in detail. Secondly, birdsong is one of the few behaviors which is performed by a separate chain of brain regions and is therefore relatively easy to investigate neurophysiologically. Work on song also provides a clear illustration of the subtle way in which birds are influenced by their internal and external environments during development.

Neurogenesis and sensitive periods in avian song learning

In many species of birds the propensity to learn songs from conspecifics is greatest during one or more distinct periods in life. These 'sensitive' learning periods, together with our detailed knowledge of the neural circuitry controlling avian song, have facilitated the discovery of radical neuroanatomical changes that accompany vocal development. One of the most remarkable of these changes is the production and incorporation of new, song-related neurons. The neurogenesis of specific cell types during song development helps create and recreate motor pathways for song production and provides synaptic plasticity that may both encourage and temporally constrain learning.

Testosterone induction of song in photosensitive and photorefractory male sparrows

Song in male songbirds is activated by the sex steroid testosterone (T). Using male song sparrows (Melospiza melodia), we compared effects of T in the normal spring state of photosensitivity (i.e., when the pituitary-gonadal axis is sensitive to stimulation by increasing daylength) and in the late summer-early fall state of photorefractoriness (i.e., when they are insensitive to increasing daylength). Photosensitive males experienced short days for 8 weeks and then long days for another 22 weeks to induce photorefractoriness. T implants were given to the birds twice, first when on short days and photosensitive, and second when on long days and photorefractory. Song rates were compared among 5 conditions: (1) photosensitive, short days, low T titers; (2) photosensitive, short days, high T titers; (3) photosensitive, long days, high T titers; (4) photorefractory, long days, low T titers; and (5) photorefractory, long days, high T titers. Plasma levels of T were monitored throughout the experiment by radioimmunoassay. T was equally effective in inducing song in both the photosensitive and photorefractory conditions. Thus, no seasonal change was found in the sensitivity to hormone action of the neural target sites mediating this behavior in song sparrows. Photosensitive birds sang at a higher rate when on long days than when on short days, however, even though there was no concomitant increase in plasma levels of T. This finding suggests that environmental factors can alter the expression of song activated by similar levels of T.

Addition of song-related neurons in swamp sparrows coincides with memorization, not production, of learned songs

During song learning in birds, neurons are added to some song nuclei and lost from others. Previous studies have been unable to distinguish whether these neural changes are uniquely associated with memorizing a song model (sensory acquisition) or vocal practice (sensorimotor learning). In this study we measured changes in neuron number within song nuclei of swamp sparrows, a species in which the two phases of song learning are nonoverlapping. Male swamp sparrows were collected as hatchlings and tape-tutored from approximately 22 to 62 days of age. Swamp sparrows memorize about 60% of their song material during this period, but do not begin practicing this learned material until approximately 275 days of age. Birds were sacrificed at 23, 41, 61, 71, 274, or 340 days of age. During sensory acquisition, neuron number increased drastically in both the caudal nucleus of the ventral hyperstriatum (HVc) and Area X. The period of sensorimotor learning was not associated with any further changes in neuron number within these regions. We were unable to detect any significant changes in neuron number within the magnocellular nucleus of the neostriatum or the robust nucleus of the archistriatum during either stage of song learning. These results raise the possibility that ongoing addition of HVc and Area X neurons may encourage, and thereby temporally restrict, song acquisition.

Estrogen stimulates the incorporation of new neurons into avian song nuclei during adolescence

In zebra finches the song control nuclei hyperstriatum ventralis pars caudalis (HVc) and area X of the lobus parolfactorius (LPO), continue to add new neurons during the juvenile period of song learning. Normally, males add many more of these new cells than do females (who do not sing), leading to pronounced sexual dimorphism within these regions. Exposing females to estradiol (E2) shortly after hatching masculinizes the HVc and area X and such females sing in response to later androgen stimulation. We investigated whether exposing female hatchlings to E2 stimulates the incorporation of HVc and area X neurons born during the juvenile period, and whether later androgen stimulation further influences addition of these late-generated neurons. Females were implanted with E2 on day 3 and received either empty or dihydrotestosterone (DHT)-filled capsules on day 25. These females, and normal males and females received [3H]thymidine daily between 20 and 40 days and were killed at 65 days. Autoradiographic analyses of HVc and area X-LPO revealed that neuron number, as well as the incidence and number of thymidine-labeled neurons was increased in E2-treated females to levels approaching those typical of males. DHT did not further influence these measures in females. These data indicate that E2 promotes either the production, migration, or survival of HVc and area X neurons born during the juvenile period.