Programmed for Preference: The Biology of Same-Sex Attraction in Rams

Abstract

The sheep is a valuable model to test whether hormone mechanisms that sexually differentiate the brain underlie the expression of sexual partner preferences because as many as 8% of rams prefer same-sex partners. This review presents an overview and update of the experimental evidence that supports this hypothesis. New evidence is presented that demonstrates a critical role for kisspeptin-GnRH signaling for regulating stable fetal testosterone levels necessary for masculinization of brain and behavior. Although these studies provide substantial support for the idea that prenatal hormones program sexual preferences, further experimentation is needed to establish causality.

Introduction

One of the most interesting topics in neurobiology relates to the question of what causes sexual orientation. The question is difficult to answer in humans owing to the complexity of our behavior and the various influences, both biological and psychosocial, which shape it. Sexual orientation can be defined as an enduring attraction to the opposite sex (heterosexuality), same sex (homosexuality) or neither sex (asexual) (American Psychological Association, 2008). Sexual attraction is the best measure of sexual orientation in humans, not sexual behavior or identity, which can both reflect social, not sexual motivations (Savin-Williams and Ream, 2007). A convergence of evidence has accumulated over the past 40 years that strongly suggests that there is a biological basis to sexual orientation in men (Bailey et al., 2016; Balthazart and Court, 2017; Roselli, 2018). This conclusion is based on correlational studies of neural traits, physical traits, twin studies, genetic and epigenetic associations, immunological studies and the effects of fraternal birth order. The most recent studies advance the idea that there may be more than one gene or mechanism that explains the development of same sex orientation in men (Ganna et al., 2019; Swift-Gallant et al., 2019). Although progress has been made to understanding a biological basis of human sexual orientation, most mechanistic investigations are limited to the study of extremely rare clinical conditions that are often confounded and difficult to interpret (Bailey et al., 2016). Thus, there is a critical need for a relevant animal model to further investigate developmental mechanisms influencing sexual orientation. Although animals do not exhibit a sexual orientation in the same sense that humans do, they do exhibit sexual partner preferences, which can be used as a reasonable approximation for human sexual orientation. Numerous rodent studies have demonstrated a developmental influence of testosterone on neural structures, sexual behaviors and olfactory preferences and have been invaluable for revealing the cellular and molecular mechanisms involved in sexual differentiation of the brain (Henley et al., 2011). However, domestic sheep have emerged as an important animal model for human sexual orientation because within natural populations of breeds common to the western US, rams can be identified that show exclusive and enduring sexual partner preference for either the same or opposite sex (Perkins and Roselli, 2007). This review will present research from my laboratory that has capitalized on this unique animal model to study potential biological mechanisms that may underlie same-sex sexual behavior in rams.

The Male Oriented Ram

Rams exhibit variability in sexual interest and sexual partner preferences (Perkins and Fitzgerald, 1992; Price et al., 1988; Roselli et al., 2004). To identify their sexual partner preferences, rams are given a series of tests (Perkins and Roselli, 2007). First, they are given tests with two sexually receptive ewes on at least nine separate occasions during which all sexual behaviors are recorded. After these initial tests, a subset of sexually active rams and sexually inactive potentially male-oriented rams are given sexual partner preference tests that present them with a choice between two rams and two receptive ewes. Rams that exclusively mount other rams are classified as male-oriented, whereas rams that exclusively mount ewes are classified as female-oriented. Accurate classifications require at least two to three breeding seasons of sexual partner preference testing starting when rams are 16–18 months old. The majority of rams are female-oriented, estimated between 60% and 70%. However, a significant proportion, approximately 8%, meet the criterion for classification as male-oriented. A male-oriented sexual partner preference can only be evaluated in the context of a choice test and differs from male-male mounting performed by rams confined in same-sex groupings. Although male-oriented rams exhibit what is considered a female-typical sexual partner preference, they do not respond to estrogen and progesterone with LH surges and female-typical receptive behaviors like females. In rodent models, perinatal hormone manipulations can induce males to exhibit female-typical sexual behaviors and olfactory preferences, but they do not lose their ability to also show male-typical sexual responses as thus exhibit an intermediate phenotype (Henley et al., 2011).

Prenatal hormones and neural correlates.

Several hypotheses have been proposed to explain the development of same-sex preferences in rams. These include effects attributed to same-sex rearing, genes, olfactory responsiveness and brain differences that are programmed by sex hormones during fetal life (Roselli and Stormshak, 2009). The most compelling evidence supports the idea that this behavior is related to structural and functional brain differences. Theories have focused on the sexually dimorphic nucleus of the preoptic area (SDN-POA) because this nucleus or nearby regions help mediate sexually-dimorphic mating behaviors in most if not all vertebrates (Henley et al., 2011). The SDN-POA was first identified in rats and is larger in males than in females (Gorski et al., 1978).

Sex differences in SDN-POA and mating behavior arise as a result of exposure to testosterone and estradiol during the perinatal critical period for sexual differentiation of the brain. The neurohormone hypothesis proposes that sexual preferences, being highly sexually dimorphic, are also influenced by the degree to which the nervous system is exposed to testosterone during the critical period (Ellis and Ames, 1987). Accordingly, the size of the SDN-POA can be used as a biomarker of perinatal T exposure and male-typical sexual differentiation. In 1991, Simon LeVay found support for the neurohormone hypothesis in humans when he reported that the third interstitial nucleus of the anterior hypothalamus (INAH3) was larger in heterosexual men than in homosexual men and women (LeVay, 1991). This finding was subsequently replicated, but the difference between heterosexual and homosexual men was not as large (Byne et al., 2001). Sheep have a homologue of the SDN-POA/INAH3, called the ovine SDN (oSDN), which comprises a dense cluster of cells in the central component of the medial preoptic nucleus and can be identified by its abundant expression of aromatase mRNA (Roselli et al., 2004). The oSDN is larger in female-oriented rams than in male-oriented rams and ewes. The difference in volume persists in adult sheep that were castrated and treated with testosterone, demonstrating that the differences in size are not affected by hormone concentrations in adult blood, but most likely established by exposure to testosterone during fetal development (Roselli et al., 2009). The demonstration that dimorphisms exist in INAH3 and oSDN in association with sexual orientation and sexual partner preference, respectively, offers persuasive evidence in support of the neurohormone hypothesis. However, neither the sheep nor human study addresses the question of whether the difference is the cause or consequence of the behavior.

This question of causality can only be addressed experimentally in an animal model. One way to establish a causal relationship is to establish whether the brain dimorphism appears prior to expression of sexual preferences in sheep. However, it is not possible to monitor the growth of the oSDN in individual female-and male-oriented rams over time as they develop in order to determine when size differences first emerge. Employing an alternative approach, we determined that a sex difference in oSDN volume appears in lamb fetuses by gestational day (GD) 135 before they are born (term = ~147 days), independently of social or sexual experiences and prior to the emergence of sexual preferences (Roselli et al., 2007). Then in a subsequent experiment, we demonstrated that the critical period during which testosterone masculinizes the oSDN occurs between GD 60 and GD 90, whereas the critical period for masculinization of the genitalia occurs earlier from GD 30 to GD 60 (Roselli et al., 2011). These results demonstrate that testosterone affects differentiation of the brain and genitals in different timeframes and could explain how hormone variations during gestation may produce rams that prefer to mate with other rams but still possess masculine genitals and other male-typical neuroendocrine and behavioral traits.

The critical period for sexual differentiation coincides with the time in gestation when testosterone is secreted by the fetal testes of males. Testosterone acts through androgen receptors or is metabolized to estradiol by aromatase and then acts through estrogen receptors. The fetal oSDN expresses the aromatase enzyme as well as receptors for androgens and estrogens (Reddy et al., 2014). Experiments performed to test the involvement of androgen receptors in sexual differentiation of the oSDN found that prenatal treatment with the androgen receptor antagonist flutamide from GD 60 to GD 90 significantly reduced the size of the oSDN in males, but did not completely feminize it (Roselli et al., 2014). These results suggest that androgen receptors mediate masculinization of the oSDN, but raise the question of why antagonism was not more effective. One possible explanation, is that flutamide treatment elicited hormone compensation through the hypothalamus-pituitary-gonadal axis. In mature males, gonadotropin-releasing hormone (GnRH) secreted by GnRH-expressing neurons in the hypothalamus stimulates gonadotrophs in the anterior pituitary to secrete luteinizing hormone (LH) and follicle-stimulating hormone (FSH). LH then stimulates Leydig cells of the testes to secrete testosterone, while FSH acts on Sertoli cells to support spermatogenesis. The production of testosterone is maintained at a relatively constant level through negative feedback on the hypothalamus and pituitary. Disruption of negative feedback, either by removing testosterone or blocking its action, results in elevated LH levels in an effort to restore testosterone. To test whether the hypothalamic-pituitary-gonadal axis is active during early gestation, male lamb fetuses were delivered on GD 85, 24 hours after the final flutamide injection. Both LH and testosterone were elevated significantly in flutamide-treated males. These results suggest that, like in adults, testosterone levels in the eugonadal male fetus are regulated through a negative feedback mechanism and LH secretion is normally suppressed by testosterone to maintain hormone homeostasis. Blocking testosterone action with flutamide interrupts negative feedback and LH and testosterone rises. The resulting elevated testosterone could act to reduce flutamide’s antagonism by competing for the androgen receptor and explain why the mean oSDN volume was intermediate in treated males. Thus, the gonadotropic axis in the second trimester lamb fetus appears to be active and comprise a classic homeostatic system that acts through negative feedback to maintain a stable and sufficient level of testosterone needed to complete sexual differentiation of the male fetus. The ability to compensate when androgen receptors are blocked may be the reason, in part, why in a subsequent experiment, we observed that prenatal treatment with flutamide did not alter the proportion of male offspring exhibiting same-sex preferences in adulthood (Roselli et al., 2016b). For this reason, the question of whether prenatal androgen plays a causal role in the development of male-typical sexual preferences remains unanswered and will require further experimentation.

A role for kisspeptin-GnRH signaling in brain sexual differentiation.

The insights gained from the flutamide studies emphasize the necessity for understanding the developmental regulation of the fetal hypothalamic-pituitary-axis in order to devise better experimental approaches that test the neurohormone hypothesis. Sheep are long gestation animals that have a long prenatal critical period that requires sustained exposure to testosterone for complete masculinization. As a general rule, testosterone secretion by the fetal testes becomes LH-dependent at the same time or shortly after the development of the hypothalamic-pituitary-gonadal axis, which occurs around GD 60 in sheep (O’Shaughnessy and Fowler, 2011). The demonstration that LH and testosterone are elevated in the serum of GD 85 male lamb fetuses after androgen action is inhibited agrees with other evidence that suggests the cellular machinery necessary for GnRH control of the pituitary is in place by the onset of the oSDN critical period (Brooks et al., 1996). The first appearance of GnRH immunopositive terminals in the external zone of the median eminence at GD 50 coincides with the completion of vascular connections between the hypothalamus and pituitary gland (Matwijiw et al., 1989) and precedes detection of serum gonadotropin on GD 55 (Foster et al., 1972). These observations beg the question of what is responsible for activating GnRH neurons in fetal lambs. In adults, kisspeptin neurons in the arcuate nucleus that co-express neurokinin B and dynorphin, i.e. the KNDy neurons, play a key role in regulating GnRH secretion and gonadal hormone feedback (Moore et al., 2018; Navarro and Tena-Sempere, 2011). Evidence from experiments with transgenic mice suggests that kisspeptin may regulate the in utero activity of GnRH neurons (Kumar et al., 2014; Kumar et al., 2015) and induce the postnatal surge of testosterone that masculinizes the brain (Clarkson and Herbison, 2016). Support has also started to accrue for the idea that kisspeptin neurons drive the gonadotropic axis and respond to negative feedback during the critical period in sheep. The mRNA for kisspeptin, neurokinin B, prodynorphin and their receptors are expressed in the fetal sheep hypothalamus as early as GD 60 (Roselli et al., 2016a). The expression of kisspeptin mRNA was significantly greater in females than in males and, together with serum LH, was reduced in females after testosterone exposure. Given this background, we set out to test the hypothesis that, in the ovine fetus, kisspeptin acts on GnRH neurons to control testosterone levels needed for sexual differentiation.

Recently, we identified kisspeptin-immunoreactive neurons in the arcuate nucleus of GD 60 and GD 85 lamb fetuses (Amodei et al., 2019). Kisspeptin neurons were more abundant in females than in males and co-expressed neurokinin B, thus providing evidence that KNDy neurons, critical regulators of GnRH secretion, exist in the sheep hypothalamus by midgestation. The sex difference most likely reflects the fact that at these ages testosterone secretion is higher in males than in females and thus provides negative feedback on GnRH and LH secretion through the inhibition of kisspeptin and neurokinin B expression. However, further studies will be needed to directly test this deduction. Studies were also performed in cannulated GD 85 lamb fetuses to determine whether activation of the kisspeptin receptor is associated with GnRH release by tracking LH after intravenous injection of the murine kisspeptin-10 (KP-10) agonist. A single bolus injection of KP-10 (10 μg) elicited a robust release of LH that was accompanied by a delayed rise in serum testosterone in males. Pretreatment with the GnRH receptor antagonist (acyline) abolished the LH response to KP-10 confirming that kisspeptin acts through the GnRH neuron. Taken together, these results support the hypothesis that kisspeptin activates GnRH neurons in male fetuses, which stimulates the pituitary to secrete LH. LH then stimulates the fetal testis to secrete testosterone, which ultimately masculinizes the brain (Fig. 1). The data also suggest that during the oSDN critical period the gonadotropic axis is established to maintain steroid homeostasis in order to ensure that sufficient testosterone exists to complete testicular development and regulate brain masculinization (Brooks et al., 1996; Roselli et al., 2011).

Fig. 1.

Diagram of the proposed fetal kisspeptin–GnRH neuron signaling pathway and feedback loop controlling male-specific testosterone secretion during the critical period for masculinization of the fetal sheep brain.

Summary and Conclusion.

The structural differences in the hypothalamus of homosexual men and male-oriented rams provide some of the most compelling evidence to date that sexual orientation is affected by the organizational actions of testosterone that are responsible for sexual differentiation of the brain. Experiments in sheep established that the oSDN develops before birth and depends on androgen receptor activation for masculinization. However, manipulations expected to alter sexual partner preferences by interfering with androgen signaling during gestation did not succeed ostensibly because the hypothalamus-pituitary-gonadal axis compensated by increasing testosterone synthesis and mitigating androgen receptor inhibition. These results demonstrate that the fetal hypothalamus-pituitary-gonadal axis is functional at the beginning of the second trimester in sheep. This is about the same time in gestation that it is activated in human males (Scott et al., 2009). For rodents, the hypothalamus-pituitary-gonadal axis is active at the very end of gestation (Kreisman et al., 2017). In all three species the time when the hypothalamus-pituitary-gonadal axis begins to function corresponds to the period in development when brain masculinization also occurs. It is interesting to speculate that this affords the masculinization process some level of protection from disruption by environmental factors that could interfere with testosterone action. Our newest results suggest that the secretion of testosterone by the fetal lamb testes in the second trimester depends on kisspeptin-GnRH signaling in the hypothalamus. Thus, we piloted the alternative approach of using treatment with a long-acting GnRH antagonist to block testosterone production in the fetus. Studies are underway to assess whether this manipulation leads to a higher percentage of rams exhibiting same-sex preferences and corresponding reductions in oSDN volumes. This will provide critical data needed to test whether a causal relationship exists between prenatal testosterone exposure and sexual partner preference.

Highlights.

• The ovine sexually dimorphic nucleus is dimorphic with respect to sexual partner preference.

• The ovine sexually dimorphic nucleus develops during a discrete prenatal critical period under the influence of testosterone.

• Kisspeptin-GnRH signaling plays a pivotal role to regulate testosterone secretion by the fetal lamb testes during the critical period.