The intricate issue of the role of aromatization for androgen actions on male sexual behaviors has been splendidly illustrated in studies of male mice with deletions of genes coding for either the estrogen a, the estrogen (3 or the estrogen a and ( receptors as well as for the androgen receptor. Mice carrying a gene deletion are normally called knockout mice and they are frequently given a first name containing the name of the gene product that has been eliminated. For example, a mouse not expressing the estrogen receptor a is called 'estrogen receptor a knockout'. Before making a brief description of the sexual behaviors in mice carrying the different gene deletions, a short note on sexual differentiation is in order. The subject of hormonal influences on morphological and behavioral development is extremely complicated and it would be utterly pretentious on my part to make believe that I could give a reasonable summary of these influences. Nevertheless, from my vantage point of profound ignorance I can allow myself to ascertain that steroid hormones from the developing testicle affect the development of the internal and external sexual organs. In the absence of a functional testicle or of functional androgen receptors, the individual will develop female external genitalia and the internal sexual organs will be abnormal. It is also very likely that fetal or perinatal steroid hormones have profound influences on the brain. In the case of a mouse or a rat, the brain will develop with the wiring typical of a female if no gonadal steroids are present during development and with the wiring typical of a male if such hormones are present. The early effects on the brain of steroid hormones are called 'organizational', in the sense that they organize the central nervous system to be like that of a male. The absence of gonadal hormones allows the brain to remain disorganized, that is to be like that of a female. I admit that this statement may give the impression of being somewhat odd, but I see no other way to interpret the term 'organizational'. Once the brain has either been organized or left disorganized, the blood concentration of gonadal hormones will be very low until puberty. At the moment of puberty, these hormones will display their 'activational'
effects. They will act on the organization prepared by the fetal or perinatal hormone secretions when stimulating sexual behaviors.
In mice with gene deletions, the organizational effects of gonadal hormones will be abnormal. The precise abnormality depends on the deleted receptor/receptors, but it is evident that the anatomical and functional substrate for gonadal hormone action in adulthood is different from that of a non-manipulated animal. An unavoidable consequence is that organizational and activational effects are confounded when studying steroid hormone actions in adult animals carrying steroid receptor gene deletions. Although extremely interesting, the results of studies employing knockout models must be interpreted with much caution because of this confound. An absent organizational effect may lead to compensatory changes in completely unrelated receptor systems, for example in neurotransmitter receptors, or changes in the connectivity of some brain areas, and other unknown changes that we cannot imagine. As a result, it is very risky to advance hypotheses concerning physiological roles of hormone receptors solely on the basis of data from knockout studies. Fortunately, in the case of sexual behaviors we have data from many other sources and, as we shall see, the data obtained in studies of knockout mice mostly coincide with earlier observations employing other techniques for blocking receptor actions.
In this context I cannot refrain from making a small comment on the popularity of knockout mice. In many situations, they are the only option for evaluating the possible functions of a gene product. In these situations, their use is beyond reproach. However, in the case of gonadal steroid receptors, the situation is entirely different. There are many well characterized pharmacological tools available if we want to block a particular steroid receptor. Eliminating the gene coding for the receptor is, in this case, a rather bad alternative. We should keep in mind that a good antagonist blocks receptor action in an efficient way and it does so at a specific time. Thus, a receptor antagonist administered to an adult animal will act in a completely normal brain and at the moment we want it to. We can even make the antagonist act at a specific site, for example by intracerebral administration of minute amounts of it. This temporal and spatial specificity of action can never be obtained by gene deletions. Thinking it over, it is difficult to understand why knockout mice have become so popular in studies of behavioral effects of gonadal steroids. They carry many disadvantages and I cannot see one single advantage with them compared to classical pharmacological procedures employing the administration of antagonists. One possible explanation is that the generation of knockout mice frequently represents an achievement of molecular biology. Because it has become a kind of fashion to profit from new achievements in that area of biology, I suspect that some scientists are irresistibly attracted to knockouts. Nevertheless, sooner or later the attraction to novelty will be replaced by a realization of the disadvantages associated with the use of these mice.
After this comment on research tactics, we can now return to studies of male mice lacking genes for one or more gonadal hormone receptors. Mice without an estrogen receptor a show a deficient sexual behavior. They may make some mounts, but they rarely perform intromissions and ejaculation is almost never seen (Rissman et al., 1997; Wersinger et al., 1997; Ogawa et al., 1998). These mice have blood androgen concentrations above normal (Eddy et al., 1996; Rissman et al., 1997), so their deficient copulatory behavior cannot be a consequence of reduced availability of androgens. The differences in blood androgen concentrations between the knockout and the wild type was, in fact, eliminated in some of the behavioral studies by using castrated males given testosterone replacement. What the data show is that the estrogen a receptor is necessary either for the organization of brain circuits involved in sex behavior or for the activational effects of gonadal hormones in post-pubertal animals or for both the organizational and activational effects of these hormones. They do not provide us with any information that we did not already have from experiments with neonatal or adult castration or neonatal or adult administration of estrogen receptor antagonists.
After elimination of the estrogen receptor (3 we see no deficit in sexual behavior (Ogawa et al., 1999). This may be a new finding, because there are no published studies of sexual behaviors in animals treated with antagonists for this receptor. In mice lacking both the a and the ( estrogen receptors, sexual behavior is completely absent. The males do not even mount (Ogawa et al., 2000). The fact that the combined deletion of the a and the ( estrogen receptor genes has more dramatic effects than deletion of the a receptor gene alone suggests that the ( receptor has some role at some time, although its absence does not by itself modify male mouse sexual behavior. Nevertheless, data from the different estrogen receptor knockout mice convincingly show that estrogen receptors are needed for the full expression of male sexual behavior. One serious limitation to these studies is that the knockout mice were all generated by gene deletions in the 129 strain and the descendants were backcrossed into the C57/BL6 strain. We know that stimulation of estrogen receptors is necessary for sexual behavior in some strains but not in others, and currently there is no way of knowing if the results obtained in the single knockout strain are applicable to other strains. Of course, it is also impossible to generalize to other species since there are so many data showing that the need for estrogen receptor stimulation in the activation of male sexual behavior is not a universal characteristic.
There are also a few studies that have looked upon sexual behaviors in mice lacking the gene coding for aromatase. This kind of knockout mouse would have much reduced estrogen concentrations in the brain and in other tissues where aro-matase normally is present. Like the estrogen receptor knockouts, the aromatase knockouts were generated in the 129 strain and backcrossed into the C57/BL6 (Honda et al., 1998). Not surprisingly, the knockout show severely disrupted cop-ulatory behavior. In addition, the males do not show any approach behaviors to estrus females or their odors in several different test procedures (Bakker et al., 2002). The impaired sexual behavior of mice lacking the aromatase gene is not surprising in view of the results from mice lacking the estrogen receptor genes previously mentioned. Although the aromatase mice may have normal amounts of estrogen receptors, these receptors should be inactive because there are no estrogens that can bind to them. Thus, it should be expected that the behavioral consequences of lack of estrogen receptors or of aromatase should be very similar, an expectation that was experimentally confirmed. While the aromatase knockout mice certainly are most interesting, the data concerning sexual behaviors obtained from them do not offer any new information. The use of a new technique has only confirmed what we already knew. As was the case with the estrogen receptor knockouts, these data are valid only for the C57/BL6 strain. Most unfortunately, they offer no evidence whatsoever as to the role of aromatization in other mouse strains or in other species.
A spontaneous mutation of the androgen receptor has been described in several mammalian species including mice and rats. It was first described in the rat in 1964, but it was not until 1973 that a complete description of it became available (Stanley et al., 1973). The animals affected were called Tfm (testicular feminized male). A single transition mutation within exon E of the androgen receptor gene is normally the cause (Yarbrough et al., 1990). An equivalent condition in mice has also been described (Lyon and Hawkes, 1970). Due to a lack of a nucleotide in exon 1 of the androgen receptor gene, the receptor does not bind to the DNA molecule and has no effect when activated by androgens (Gaspar et al., 1991; He et al., 1991). Tfm rats and mice are identical to the human in the way that genetic males have female external genitalia and functional, non-descended testes. Many endocrine and other kinds of physiological studies have been performed on these animals, but there are few studies of sexual behavior (Olsen, 1979, 1992). Nevertheless, the existing studies show that male sexual behavior is severely impaired in tfm rats and mice, independently of whether the animals are castrated or not, and independently of adult hormone treatments (Ohno et al., 1974). The data from studies on this natural mutation are unequivocal in the way that they clearly establish that androgen receptors are necessary for masculine sexual behavior in rats and mice.
This idea is further strengthened by a study employing mice in which the androgen receptor gene was artificially altered. The appearance of these male mice was that of a female, exactly as is the case in the natural testicular feminization syndrome. Furthermore, they displayed no sexual behavior at all, neither when intact nor when castrated and treated with dihydrotestosterone. All the wild type littermates mounted, intromitted and ejaculated when intact and displayed enhanced copulatory behavior after treatment with dihydrotestosterone (Sato et al., 2004). The androgen receptor knockout mice were generated from embryos of the CD-1 strain and floxed C57/BL6 mice were backcrossed with CMV-Cre transgenic mice, according to the methods description. This elegant procedure is certainly comprehensible to some molecular biologists, but I must admit that the only thing clear to me is that the 129 strain was not involved. I mention this only because strain may be, as we already know, of crucial importance in studies of mouse behavior. Independently of this, and also taking into account the data from animals with the androgen insensitivity syndrome, we can probably accept the conclusion that a functional androgen receptor is necessary for male sexual behavior in at least some strains of mice and rats.
Was this article helpful?