Defending the Brain from Estrogen

NEWS AND VIEWS

Defending the brain from estrogen

David A Puts, Cynthia L Jordan & S Marc Breedlove

Alpha-fetoprotein binds estrogens in the developing brain. A new paper shows that inhibiting estrogen rescues the brain masculinization found in female mice lacking this gene, suggesting that alpha-fetoprotein inhibits estrogen activity in females.

In some circles, it is popular to blame ‘testos- son, developing females need to be protected males it would leave testosterone free to terone poisoning’ for the aggressive tenden- both from estrogens, which can directly enter the brain and be locally aromatized cies and indiscriminate sexual interests of masculinize the brain, and from testoster- to estrogen to masculinize and defeminize men, but evolutionary theorists insist such one, which can be converted to estrogens. the male’s brain. In females, though, the characteristics are adaptive male reproduc- The most compelling demonstration of how AFP binding of estrogens could keep them tive strategies. Thus, in many mammalian early estrogen exposure can affect females out of the brain, circulating long enough to species, prenatal exposure to testosterone is that a single dose of estrogen given to a be metabolized into inactive steroids. This and other androgens induces the male, in newborn female rat will prevent her brain vision of brain sexual differentiation com- adulthood, to behave in ways that help him from maintaining ovulatory cycles4. What’s fortably explained most reports in the litera- compete with other males for mates. But more, even if such a female is treated in ture, and seemed to resolve the question of what is good for the gander may not be adulthood with the steroid hormones that how estrogens could masculinize the male good for the goose: an aggressive response normally induce sexual receptivity, she will brain but not the female brain. to males is maladaptive for female repro- not respond5. When mounted by a male, As is usual with a new hypothesis, find- http://www.nature.com/natureneuroscience ductive success. Instead, the female must, at she will not display the lordosis posture ings soon arose that were inconsistent with some point, be sexually receptive to males if (raised head and rump, deflected tail) that the idea that AFP acts by keeping estrogen she is to reproduce. So for females, testos- permits the male to insert his penis into her out of the brain. For one thing, several terone exposure during development might vagina. Such a female is called ‘defeminized’ studies suggested that removing the ovaries really constitute poisoning. How do females because she is less likely to display female- from newborn rats made their behavior avoid testosterone exposure, especially in typical behaviors such as lordosis. She may less feminine8–12, not more feminine as one mammalian species with large litters where also be ‘masculinized’: that is, more likely to would predict if ovarian estrogens masculin- almost every female fetus shares the uterus display male-typical behaviors. For example, ize the brain. These findings are somewhat with brothers? In this issue, Bakker et al.1 as adults, such females will be more likely to inconsistent across studies, indicating that examine knockout mice to demonstrate mount other females. In any case, perinatal the feminizing effect of early estrogens, if how a particular protein prevents steroid exposure of the female brain to estrogen real, is a subtle one. Because we know that hormones from masculinizing the brains of severely disrupts her chances to reproduce. a lot of estrogen will defeminize the brain,

Nature Publishing Group Group 200 6 Nature Publishing developing females. Fetal females are protected from testoster- the question is whether a very modest dose

© To understand their study, one must one itself by the placenta. Because it is rich of estrogen is needed to complete feminine first realize that in rats and mice it is not in aromatase, the placenta readily converts development. testosterone itself that masculinizes the androgens passing through, such as testos- Such findings led Dominique Toran- developing brain, but rather metabolites terone, into estrogens6. This means that little Allerand of Columbia University to suggest of testosterone: estrogens such as estradiol testosterone from mouse brothers in utero that perhaps the role of AFP was to bring masculinize many brain regions2. The fetal is likely to reach their sisters. Obviously, estrogen into the brain13, a suggestion testes produce androgenic steroids, such as though, this mechanism does nothing to strengthened by her immunocytochemical testosterone, that themselves masculinize shield females from estrogens, which mol- detection of AFP inside brain neurons14. many body structures. In the brain, however, ecule-for-molecule are more effective than What’s more, the AFP does not seem to be testosterone is converted in a single step to testosterone at masculinizing the developing made in the brain, so it is being transported estradiol. This conversion is accomplished brain. In addition to aromatized metabolites into neurons from the outside. Because AFP by the enzyme aromatase, and the develop- of testosterone from her brothers, the female binds estrogens, perhaps it is escorting estro- ing brain has loads of aromatase on hand for fetus is also exposed to estrogens from the gen into the brain to favor feminine behav- the task3. The locally produced estrogen then maternal ovary and possibly her own. What ioral development (Fig. 1b). Presumably this binds with estrogen receptors (not androgen keeps all this circulating estrogen from mas- effect of AFP-escorted estrogen would result receptors) to modulate gene expression and culinizing the female brain? in very selective and specific estrogen deliv- sculpt developing brain circuits into a mas- Several decades ago, Bruce McEwen of ery to particular sets of neurons equipped culine configuration (Fig. 1a). For this rea- Rockefeller University proposed that a cir- to internalize the AFP-steroid complex. It culating protein, alpha-fetoprotein (AFP) is easy to imagine that selective estrogen might do the trick7. As the name suggests, delivery to a specific neural circuit might this protein is found in the circulation of contribute to feminine brain development. The authors are in the Neuroscience Program, fetuses, and in rats and mice AFP binds a Because it was difficult to identify which Michigan State University, East Lansing, Michigan wide variety of steroid hormones, includ- particular neurons are benefiting from this 48824-1101, USA. ing estrogens. Conveniently enough, AFP estrogen stimulation, it was difficult to dis- e-mail: [email protected] does not bind testosterone very well, so in prove definitively either the hypothesis that

NATURE NEUROSCIENCE VOLUME 9 | NUMBER 2 | FEBRUARY 2006 155 NEWS AND VIEWS

T Testosterone Female Female abMale E cE αFP αFP T T

E T E αFP αFP Aromatase

Cytoplasm E Estrogen

E Moderate E Little E Estrogen ER ER activation ER ER activation ER receptor

E E E Nucleus ER ER ER

Defeminized Altered gene Feminized Altered gene Feminized Altered gene

brain expression brain expression brain expression Debbie Maizels

Figure 1 Sexual differentiation of the brain. (a) In male rodents, testosterone reaches the fetal brain and is aromatized there into estrogens (E), which bind to estrogen receptors (ER) to promote gene expression that masculinizes some neural circuits and defeminizes others. There remained the question of whether http://www.nature.com/natureneuroscience optimal development of the female brain is best served by delivery of some estrogen to specific neural circuits or by protecting the brain from estrogen as much as possible. These competing hypotheses suggested different roles for alpha-fetoprotein (AFP), which binds estrogens. (b) AFP might deliver estrogen to specific neural elements to promote feminization of those circuits. (c) Alternatively, AFP might serve to keep estrogens out of the brain of fetal females. Females lacking the Afp gene were found to be defeminized and masculinized, as both hypotheses would predict. However, blockade of estrogen synthesis in the Afp–/– females restored feminine behavior, which is compatible only with the hypothesis in c.

AFP keeps estrogen out of the brain (Fig. 1c) the two competing hypotheses, these out- mouse behavior. The protein is not a chaper- or that it selectively brings estrogen into comes are consistent only with the notion one, ushering estrogen into the female brain, the brain. that AFP functions by excluding estrogen but more like a matron, firmly excluding For example, in the mice studied by from the female’s brain. estrogen from the developing brain so that 1

Nature Publishing Group Group 200 6 Nature Publishing Bakker et al. that have the gene for AFP What’s next? This finding does not explain females can feminize themselves, untainted –/– © knocked out (Afp mice), both hypotheses why neonatal ovariectomy sometimes inter- by steroidal interference. would predict that the males would be more feres with feminine development. Perhaps or less normal but that the females would be there is a mechanism for selectively deliver- defeminized and masculinized. That is what ing a modicum of estrogen to a developing 1. Bakker, J. et al. Nat. Neurosci. 9, 220–226 –/– (2006). was found: Afp females treated as adults feminine circuit, but AFP just does not hap- 2. Morris, J.A., Jordan, C.L. & Breedlove, S.M. Nat. with hormones that normally induce sexual pen to be that mechanism. Likewise, we are Neurosci. 7, 1034–1039 (2004). receptivity show very little lordosis response still left with the question of why AFP seems 3. Roselli, C.E. & Resko, J.A. Steroids 50, 495–508 (1987). to mounting males (they are defeminized) to be internalized into neurons in the devel- 4. Gorski, R.A. Am. J. Physiol. 205, 842–844 and are more likely to mount other females oping brain. Perhaps AFP is escorting some (1963). (they are masculinized)1. By themselves, other player into neural circuits. It is clear 5. Whalen, R.E. & Nadler, R.D. Science 141, 273–274 (1963). these results cannot tell us whether the that such putative escorting of ‘factor X’ is 6. Conley, A. & Hinshelwood, M. Reproduction 121, absence of AFP defeminized and masculin- not required for females to display lordosis 685–695 (2001). –/– 7. McEwen, B.S., Plapinger, L., Chaptal, C., Gerlach, ized the brain by allowing estrogen in or by because Afp mice, when protected from J. & Wallach, G. Brain Res. 96, 400–406 (1975). not bringing estrogen to the select neuro- estrogens perinatally, were sexually recep- 8. Blizard, D. & Denef, C. Physiol. Behav. 11, 65–69 nal circuit. However, treating the developing tive as adults, but it may well be involved in (1973). –/– 9. Sodersten, P. J. Endocrinol. 70, 409–420 (1976). Afp females with a drug that systemically some other brain function. Perhaps a closer 10. Gerall, A.A., Dunlap, J.L. & Hendricks, S.E. J. Comp. blocks estrogen synthesis dramatically look at the neurons that contain AFP will Physiol. Psychol. 82, 449–465 (1973). rescued the feminine phenotype. In other suggest a function for the protein in these 11. Leret, M.L., Molina-Holgado, F. & Gonzalez, M.I. Physiol. Behav. 55, 371–373 (1994). words, by reducing the amount of estrogen cells. If so, it could be a function common 12. Denti, A. & Negroni, J.A. Acta Physiol. Lat. Am. 25, to which the fetuses were exposed, the inves- to both sexes, as males also produce AFP. It 99–106 (1975). –/– –/– 13. Toran-Allerand, C.D. in Sex Differences in the Brain: tigators allowed Afp females to display will be interesting to see whether these Afp the Relation Between Structure and Function (eds. sexual receptivity again. Blocking perinatal mice have other, unanticipated differences De Vries, G.J., De Bruin, J.P.C., Uylings, H.B.M. & estrogen synthesis also restored female levels in behavior compared to wild-type mice. Corner, M.A.) 63–98 (Elsevier Science, Amsterdam, 1984). of tyrosine hydroxylase–expressing neurons In the meantime, we now know that AFP is 14. Toran-Allerand, C.D. Brain Res. 281, 213–217 in the hypothalamus of Afp–/– females. Of crucial indeed in sexual differentiation of (1982).

156 VOLUME 9 | NUMBER 2 | FEBRUARY 2006 NATURE NEUROSCIENCE