Hormones and the development of sex differences in behavior
- First Online:
- Cite this article as:
- Adkins-Regan, E. J Ornithol (2007) 148: 17. doi:10.1007/s10336-007-0188-3
- 129 Views
Birds exhibit striking diversity in behavioral sex differences. A necessary complement to the study of the ecology and evolution of these sex differences is discovering the proximate physiological mechanisms for their development (sexual differentiation) and adult expression. Experiments with Japanese quail (Coturnix japonica) have shown that sex differences in crowing, strutting, and sexual receptivity are produced partly or largely by hormonal dimorphism in adulthood (activational effects of sex steroid hormones), whereas the sex difference in copulatory mounting is produced by permanent actions of sex steroids occurring early in development, during the embryonic period in this precocial species (organizational effects). Experiments with zebra finches (Taeniopygia guttata), an altricial species, have revealed organizational effects on singing and mating that occur after hatching. In both species, sex differences in whether birds are interested in females vs. males are produced by organizational rather than activational effects. Results of experimental manipulations of sex steroid actions in juvenile zebra finches suggest hormonal regulation of the onset of interest in the opposite sex as sexual maturity is reached. Zebra finches are socially monogamous and permanently paired across breeding attempts. Experimental reduction of sex steroid actions had no effect on pairing success in either sex. The regulation of adult pair formation by sex hormones is more likely to occur in species that pair seasonally. The concepts of organization and activation and the results of these experiments raise a number of questions and are a potential source of hypotheses about developmental changes responsible for the evolution of species diversity in sex differences.
KeywordsJapanese quailPair formationSex steroidSexual differentiationZebra finch
The roughly 9000 species of living birds exhibit an impressive diversity with respect to behavioral differences between the sexes. Such differences are quite marked in species such as turkeys, manakins, and ruffs, in which males but not females engage in elaborate courtship displays, whereas females but not males perform all of the parental care. In buttonquail, the roles are reversed, and females but not males produce advertisement calls (Madge and McGowan 2002). In albatrosses and many other seabirds, on the other hand, there are few obvious sex differences in behavior, and displays tend to be performed by both sexes together.
Understanding both the fact of sex differences and the diversity in their nature and magnitude can be achieved by addressing all four of Tinbergen’s (1963) aims for ethology: function, phylogeny, development, and physiology. These last two have been the focus of this author’s research program. The progress that has been made by the author and other researchers and the many unanswered questions that remain form the subject of this review.
Sexual differentiation of behavior: conceptual background
Sexual differentiation is the developmental process that produces sex differences in morphology and behavior. Although birds have sex chromosomes, most of the genome is in the autosomes. The fundamental question then is, how are two sexual phenotypes created from a largely monomorphic genome? There are several possibilities (reviewed in Adkins-Regan 2005). The small portion of the genome that is dimorphic (the genes on the sex chromosomes) could be directly responsible. Alternatively, sex differences in behavior could be learned through experience or by observing others. They could also be situational, arising from local demography or social context (for example, group sex ratio), and could reflect decisions about the appropriate tactics to adopt in that context. They could arise through hormone action. In 1959, Phoenix, Goy, Gerall, and Young proposed that organization very early in development by an animal’s endogenous sex hormones was the mechanism producing the sex difference in the mating behavior of guinea pigs. Their theory remains an essential conceptual starting point for an understanding of how hormones contribute to different behavior in males and females. They distinguished between two kinds of effects of sex steroid hormones: activational and organizational. Activational effects occur at puberty and in adulthood and serve to regulate behavior, allowing it to be expressed when that is appropriate. Typical examples would be the cyclic estrous behavior of female mammals, which is activated by ovulatory levels of estrogens or estrogens plus progesterone, and seasonally occurring breeding behavior (territory establishment, nest building, mating) of birds living in the temperate zone, which is activated by increases in gonadal androgens and estrogens (Wingfield and Farner 1993). Organizational hormone effects, on the other hand, occur during a limited critical period early in development, the fetal or neonatal period in the case of mammals. During this period sex steroids produced by the gonads permanently establish the future behavioral phenotype of the animal (masculine or feminine).
According to this theory, sex differences in the behavior of adults can result from either activational or organizational effects. If the adult sex hormone levels are sexually dimorphic, and if sex reversal of those levels results in sex reversal of behavior, one would conclude that the sex difference was activational. For example, in a number of species of birds, normally only males sing, but the administration of testosterone to adult females causes them to begin singing (reviewed in Balthazart and Adkins-Regan 2002). The females are capable of singing but do not normally sing because they lack sufficient activational testosterone. If sex reversal of adult hormone levels does not alter the behavior, however, then an organizational hormone effect that occurred much earlier in development is an alternative hypothesis. As the Phoenix et al. (1959) experiment and numerous others have shown, many behavioral sex differences in mammals are produced by hormonal activation, organization, or a combination of both, and organization follows a pattern in which the presence of testicular hormones in early development masculinizes the nervous system of genetic males and the absence of testicular hormones produces a feminized nervous system (Nelson 2005). In this manner, two sexual phenotypes are created from an initially undifferentiated embryo and a largely monomorphic genome, and the role of the sex chromosomes is mainly to establish the sex of the gonads. Although there are several other routes to adult sex differences in behavior, it is remarkable how many of those studied in mammals turn out to be produced through hormonal organization and/or activation.
With hormonal organization, sex differences can be generated that do not risk the appearance of unwanted traits in the “other” sex because the trait can be permanently suppressed or prevented from developing in one sex. This is important for understanding the evolution of traits that make a positive contribution to fitness in one sex but have a negative effect in the other. With activation alone as a mechanism, there would always be the possibility of the trait appearing to a slight degree in the “wrong” sex, because both sexes produce the same hormones, only in different quantities.
Sexual differentiation in Japanese quail (Coturnix japonica)
Does the organizational hormone theory of sexual differentiation apply to the behavior of birds? The domestic Japanese quail has been the subject of a number of experiments to answer this question. These birds have significant advantages for such research, including high fecundity, rapid development, and some clear sex differences in hormonally regulated behavior. As is typical of some other galliform birds, males emit frequent loud crows, perform a strutting display, and engage in copulatory mounting with females. Females do not show any of these behaviors and instead may adopt a receptive posture in response to male’s mating attempts. All of these behaviors are dependent on gonadal hormones for their expression. They are absent in birds that are gonadectomized or housed on short days to produce gonadal regression, and they are restored by treating males with testosterone or females with estradiol (Beach and Inman 1965; Adkins and Adler 1972; Balthazart et al. 1986).
Testosterone and estradiol levels of adult birds are sexually dimorphic, with higher testosterone in males and higher estradiol in females (Balthazart et al. 1986). Giving females testosterone elicits some crowing and strutting, and giving males estradiol causes them to adopt the receptive posture to other males’ mating attempts (Adkins and Adler 1972; Adkins 1975; Schumacher and Balthazart 1983; see also Yazaki et al. 1999). Thus, the sex differences in crowing, strutting, and receptive behavior are partly or largely (receptivity) activational in origin.
Both of these dramatic sex-reversing effects (elimination of mounting in males by embryonic treatment with estradiol and the appearance of mounting in females by embryonic treatment to prevent estrogen action) have also been obtained in the laboratory of Jacques Balthazart, where impressive progress has been made in exploring the neural bases for such effects (Balthazart and Adkins-Regan 2002). Embryonic exposure to estrogen has also been shown to reduce mounting in male chickens (Wilson and Glick 1970; Sayag et al. 1989). Taken together, these results indicate that the sex difference in these galliform birds in copulatory mounting is hormonally organized and suggest a pattern of sexual differentiation in which embryos develop as males unless exposed to estrogen before hatching, and embryos exposed to estrogen develop as females. In other words, the ovary establishes the female phenotype through an estrogenic signal, and the absence of this ovarian signal results in males. This is the opposite pattern from mammalian sexual differentiation (in which the presence of testes imposes maleness), but it is entirely consistent with findings that female embryos produce more estradiol than male embryos and have higher estradiol to testosterone ratios (Tanabe et al. 1979; Schumacher et al. 1988; Ottinger et al. 2001). Although the pattern of differentiation is reversed from mammals, the fundamental importance of hormonal organization for generating a sex difference in mating behavior is shared with mammals.
Sexual differentiation in zebra finches (Taeniopygia guttata)
The other avian species that has figured prominently in studies of how sex differences in behavior are produced is an Australian passeriform, the zebra finch, in the family Estrildidae. Zebra finches breed well in captivity or under domestication and share with Japanese quail the advantage of rapid development. Developmental mode is altricial, however, making the early posthatching period a likely time for any organizational critical period. Although zebra finches are socially monogamous in their mating system, nonetheless there are pronounced sex differences in singing (only males sing) and copulatory mounting (seldom seen in females). Adult males have higher testosterone levels than females, but there is little sex difference in estradiol levels (Adkins-Regan et al. 1990). Adult females given testosterone do not sing or show copulatory mounting, as if these sex differences have been organized earlier in development (Arnold 1980; Adkins-Regan and Ascenzi 1987).
These nestling hormone effects on adult copulatory mounting and singing support the hypothesis that the sex differences in these behaviors are hormonally organized and point to an estrogenic signal as the key for sexual differentiation in this species. However, the fact that the same treatment feminizes (demasculinizes) one behavior (mounting) while masculinizing another (singing) is quite odd and difficult to reconcile with any simple pattern of sexual differentiation. Subsequent research produced additional results that are incompatible with either a mammalian or a quail model of sexual differentiation (reviewed by Arnold and Schlinger 1993; Wade 1999; and Balthazart and Adkins-Regan 2002). The measurement of sex steroid levels during zebra finch development did not reveal any replicable sexual dimorphism in hormone levels (any dimorphic endocrine signal for sexual differentiation). In addition, the treatment of nestlings with estrogen receptor antagonists or estrogen synthesis inhibitors did not produce the effects that would be predicted; they neither masculinized copulatory mounting in females nor eliminated singing in males. Sex reversal of the gonads by treatments of embryos failed to sex reverse singing, which instead was always consistent with the bird’s genetic sex. This would seem to rule out organization by gonadal hormones altogether. And yet treatment of the nestlings with estradiol does produce singing females. What is the real signal for sexual differentiation that estradiol is perturbing, and where is it produced? The problem of explaining zebra finch sexual differentiation has not yet been solved, but alternative hypotheses are being actively pursued. One possibility is that the signal is a neurosteroid – that is, a steroid produced de novo in the brain itself in greater amounts in one sex (Schlinger and London 2006). Another is that brain sex is produced by cell-autonomous products of genes on the sex chromosomes (Arnold 2002; Wade et al. 2005). These genes could be either on the Z chromosome, with a double dose producing a male brain, or on the W chromosome, preventing development of a male brain.
Development of mate preferences
Behavioral sex differences in birds include not only motor patterns like singing and mounting, but also mate preferences. For example, females may prefer males performing elaborate displays or singing a large repertoire of songs, whereas males of these species cannot have this as a mate choice criterion if females do not display or sing. Female zebra finches prefer males with higher song rates and songs learned correctly from an adult male tutor, whereas males prefer fecund females, those that have been eating a diet likely to lead to more eggs (Williams et al. 1993; Collins et al. 1994; Jones et al. 2001). Males of many species tend to be less discriminating (choosy) in their choice of mating partners. Treatment of adult female dark-eyed juncos (Junco hyemalis) with testosterone lowers (masculinizes) their choosiness scores, suggesting the fascinating possibility that this sex difference might have an activational hormone basis (McGlothlin et al. 2004). The most obvious sex difference in mate preference is that females prefer males for mating and pairing and males prefer females. In sexual partner preference (as this phenomenon is called), it is the targets for the animals’ behavior (their approach and attraction) that differ according to the sex of the subjects and define the phenomenon – and not the behavior itself. Is sexual partner preference in birds an organizational or an activational phenomenon? In mammals, organizational hormone effects are key (Adkins-Regan 1988; Baum 2006).
An interesting study on Japanese quail by Balthazart et al. (1997) can be interpreted as supporting an organizational hypothesis for sexual partner preference in a galliform species. Male, but not female, quail are motivated to look at a female through a window, and giving females testosterone does not make them interested in other females. Females hatched from eggs injected with an estrogen synthesis inhibitor, however, are fully masculinized with respect to interest in looking at females, just as they are quite masculinized in their copulatory mounting.
To date, sexual partner preference is the only sex difference in mate preference for which an organizational hypothesis has been tested. In both Japanese quail and zebra finches, early hormone treatments can produce permanent sex reversal of this preference. In the highly social zebra finch, there seem to be contributions of some kind of social experience to this developmental process as well. In fact, manipulation of social experience alone can eliminate sexual partner preference, producing males and females that are equally interested in pairing with both sexes (Adkins-Regan and Krakauer 2000).
This conclusion about the organizational hypothesis has to be qualified, however. As with the studies of copulatory mounting and singing, there are still some unsolved puzzles. How can both estrogen and an estrogen synthesis inhibitor produce the same effect (females that prefer other females)? And if estrogen is the signal for masculinization, why does early treatment with an estrogen synthesis inhibitor fail to produce males that prefer other males (Adkins-Regan et al. 1996)?
Formation of pair relationships
The vast majority of avian species are socially monogamous, and male-female pairs are a conspicuous feature of bird life. Much research has been devoted to understanding the function and ecology of monogamy and deviations from genetic monogamy in the form of extra-pair matings. The development and physiology of pairing has been comparatively neglected. With respect to sex differences, pairbond behavior is relatively monomorphic. Zebra finches are a representative example: paired birds clump together (sit in direct physical contact), preen each other (allopreen), and spend time in the nest together, particularly during the nest building period. Nor is there any obvious sex difference in the motivation to pair. Is such monomorphic behavior even hormone-dependent? Are high (reproductive) levels of sex hormones required for pair formation and maintenance? The answers to these questions are far from obvious because this species is continually paired even when not actively breeding, and the behaviors themselves can be shown by young juveniles with their family members (Zann 1996).
This species is relatively aseasonal, breeding in response to unpredictable rainfall (Zann 1996). Such a lack of any significant involvement of sex hormones in adult pair formation may generalize to other species that are permanently paired even when not actively breeding. It is quite possible that sex hormones have a greater role in birds that form pairs on a seasonal basis, a hypothesis that needs to be tested with an experimental manipulation. Even in seasonally pairing birds, however, high levels of sex hormones do not seem to be required for continued maintenance of the existing pair relationship through the breeding cycle because sex steroids often fall dramatically at the onset of incubation, even when the pair remains together to share the incubation and feeding of the young (Wingfield and Farner 1993).
Implications and conclusions
These experiments with Japanese quail and zebra finches show that some of the elements of the theory of hormonal organization can account for the development of sex differences in the behavior of birds. In Japanese quail, it is clear that actions of estrogen during the embryonic period produce those components of the sexual phenotype that cannot be sex reversed by adult hormone manipulations. The signal that sends males and females down different developmental paths is ovarian estrogen and the estrogen to androgen ratio. In zebra finches, it is less clear what the signal is (whether it is hormones, neurosteroids, or cell-autonomous gene products), but its actions are occurring early in development and are permanent.
Many questions have not yet been answered. There are, of course, the puzzles of zebra finch sexual differentiation as well as a lack of knowledge of how sexual differentiation of behavior occurs in birds other than quail, chickens, and zebra finches. What other clades use the quail/chicken rule? Does the sister clade to the galliforms, the anseriforms, use the same rule? Whatever the zebra finch rule is, will it apply only to estrildid finches, or will it explain sexual differentiation in oscine passerines more generally? With respect to pair formation, what is the physiology of pairing in zebra finches, if it does not lie in sex steroids? What, if any, is the involvement of other molecules known to be important for affiliative relationships in mammals, particularly neuromodulators and neurotransmitters such as oxytocin family peptides, endogenous opiates, and dopamine? With respect to sexual partner preference, is the developmental process that produces preferences for the opposite sex for mating and pairing purposes the same process that produces “preferences” for the same sex for aggressive purposes? Birds often show sex-specific aggression in defending a territory or a mate, with females targeting other females and males targeting other males, a phenomenon that can be captured with laboratory testing paradigms (e.g., Adkins-Regan and Robinson 1993). Like sexual partner preference, this must involve sex recognition, and could be an organizational phenomenon.
The progress that has been made in understanding sexual differentiation opens up the possibility of understanding species differences in dimorphism and social systems with respect to the developmental changes that underlie evolutionary divergence and changes in state. Three examples illustrate some of this promise of this knowledge. First, as progress is made in figuring out zebra finch song differentiation (i.e., how male and female brains for singing are produced), testable hypotheses can be proposed to explain the developmental mechanisms for evolutionary changes from singing by males only to singing by both sexes or vice versa. Such changes have occurred a number of times in oscine passerines (Garamszegi et al. 2006; MacDougall-Shackleton and Ball 1999); for example, a change from singing by males alone to duetting has occurred in the wrens of the Thryothorus genus (Farabaugh 1982; Brenowitz and Arnold 1986; Levin 1996; Nealen and Perkel 2000; Mann et al. 2006). Second, evolutionary changes have occurred in sex roles in birds such as phalaropes, with females competing for males and males carrying out all the incubation. The hypothesis that these “reversed” sex roles are based on adult (activational) hormone sex reversal has been tested and rejected, leaving hormonal organization as one – yet untested – alternative hypothesis (Schlinger et al. 1989; Fivizzani et al. 1990; Eens and Pinxten 2000). Third, many galliform birds are polygynous with female-only parental care, but several of the species in the family Cracidae are socially monogamous with biparental care (del Hoyo 1994). The quail experiments offer the promise of accounting for reductions or increases in sex differences in parental behavior in galliform birds.
The quail experiments have also established this species as the leading model for testing the effects of endocrine-disrupting chemicals on the developing brains and behavior of birds (Halldin 2005; Ottinger et al. 2005). The sensitivity of the male embryo’s brain to estrogens means that even weakly estrogenic chemicals could seriously interfere with future copulatory behavior. Without copulation, there is no fitness. It is urgently important to know how widely this deleterious effect generalizes among birds.
The author’s research was supported by the U.S. National Science Foundation. The contributions of the following people to the research reviewed here are gratefully acknowledged: Norman Adler, Mary Ascenzi, Sunayana Banerjee, Tim Van Deusen, James Goodson, Alan Krakauer, Cary Leung, Viveka Mansukhani, Mary Ann Ottinger, Tiffany Robinson, Cynthia Seiwert, Maggie Smith, Richmond Thompson, Michelle Tomaszycki, Juli Wade, and Sharlene Yang.