Sexual differentiation of the vertebrate brain and behavior is mediated in large part by gonadal steroid hormones acting in early development. In mammals, sexually differentiated hormone production begins at the earliest stages of gonadal development and produces marked sexual dimorphism in brain and genitals by early post-natal life (for review, see [1, 2]).

Naked mole-rats (Heterocephalus glaber) do not appear to fit this framework of sexual differentiation: most individuals fail to show traditional sex differences in neuroanatomy or behavior, although these features vary based on reproductive and social status [3–6]. Naked mole-rats are eusocial rodents, living in large colonies with a strict reproductive and social hierarchy: a single female and 1–3 males reproduce and are socially dominant, while others in the colony are subordinate and usually remain reproductively suppressed for the duration of their life [7–9]. The social status within a colony is remarkably stable; however, subordinates can achieve breeder status after loss of a breeder due to death or removal from the colony. In their natural environment, subordinate naked mole-rats can—though rarely—leave their natal colony to find an opposite sex partner and establish their own colony [10, 11]. In captivity, subordinates only undergo sexual maturation if they are separated from their colony [5, 12, 13] or if the established breeders die or are removed [14, 15].

Regardless of the specific circumstances in which subordinates become breeders, they only perform sexual behaviors and show mammalian typical sex differences in gonadal steroid hormones as they transition to breeding status. Subordinate females have lower urinary and plasma progesterone (P) levels as well as plasma luteinizing hormone (LH) concentrations compared to breeder females [16–18]. Within 3 days of separation from their natal colony, subordinate females show increases in urinary P levels, and within 7 days, they develop a perforated vagina [16]. Similarly, male subordinates have lower urinary testosterone (T) compared to male breeders in the colony [14, 15, 19], and urinary T levels and plasma LH increase when subordinate males are removed from their colony [20].

Subordinates and breeders also differ in the neural structures associated with sexual behaviors. However, where sex differences in the nervous system are found in most mammals, status differences are more prominent in the naked mole-rat. For example, breeders, regardless of sex, have larger regional volume in various reproductively relevant hypothalamic structures [4], as well as increased motoneuron soma size in Onuf’s nucleus [3]. Furthermore, androgen receptor (AR) and oxytocin immunoreactivity are influenced by social status more than sex [6, 21, 22].

Collectively, these data reveal a complex interplay between sex and social status in neuroendocrine function in these highly social mammals. To investigate how an apparent reduction or lack of neural sex differences can be associated with typical sex differences in gonadal function seen in breeders, the present study evaluated the time course of changes in gene expression of gonadal steroid hormone receptors and aromatase in brain, as well as associated circulating hormones. Thus, we quantified diencephalic levels of messenger RNA (mRNA) expression of androgen receptor (Ar), estrogen receptor alpha (Esr1), progesterone receptor (Pgr), and aromatase (Cyp19a1) and measured circulating T, 17β-estradiol (E2), and P, as animals transitioned from subordinate to breeding status. We hypothesized that sex differences in steroid hormone receptors would be associated with the differences in circulating gonadal steroids that emerge in breeders, specifically predicting that females would show increases in Esr1 and Pgr while males would show increases in Ar.

The present study demonstrates that sex-specific changes in neuroendocrine function are associated with the transition to reproductive status in naked mole-rats. We observed the emergence of latent sex differences in diencephalic gene expression and confirmed the emergence of sex differences in circulating gonadal steroids within a week after the animals are removed from reproductive suppression in their home colony. As discussed below, many of these neuroendocrine changes emulate typical sex differences observed in other species during puberty and reproduction, and these results may indicate the endocrine underpinnings of eusociality in naked mole-rats.

The first change we observed in the males was an increase in Ar mRNA expression that mirrored increased circulating T at 1 week post-removal from the colony. The increase in circulating T at this time point is consistent with previous reports in male naked mole-rats [12, 20] and suggests that T action via hypothalamic Ar is associated with the onset of sexual maturation in male naked mole-rats, analogous to other mammalian species. For example, in hamsters, rats, and ferrets, circulating T at puberty is required for male-typical sexual behavior, as well as to suppress female-typical sexual behavior (for review, see [26]). Interestingly, Ar mRNA was not consistently high after removal from the colony. At 1 month, Ar mRNA levels returned to baseline subordinate levels. Transient increases in Ar expression may reflect the termination of an initial phase of reproductive maturation in the breeder males. Alternatively, or additionally, reduced AR protein may be associated with a transition into a parental state. In birds, fish, humans, and other mammals, fathers-to-be exhibit decreases in T prior to the birth of their offspring [27–32], and we have previously reported reduced AR immunoreactivity in the breeding males compared to the subordinates [21, 22]. Decreases in T have been associated with the nurturing aspects of parenting rather than the defense of young/competitive aspects of parenting (for review, see [33]). However, none of our 1-month pairs were obviously pregnant at the time of tissue collection, and T levels remained high in the 1-month paired males. Future experiments evaluating behavior during peaks in Ar mRNA, as well as manipulating circulating T levels, will test both the functional significance of changes in gene expression and the consistency across species of the association between androgens and sexual maturation/mating behavior.

It warrants mention that the status difference in Ar mRNA expression we report here, with the breeding males having a 2.5-fold increase over the subordinate males, is in the opposite direction of status differences in AR protein that we have previously reported [21, 22]. Indeed, increased AR immunoreactivity would be predicted in breeders as androgens typically both stabilize AR protein and facilitate immunohistochemical detection by causing nuclear translocation. Three possibilities might explain this apparent discrepancy. First, our work with AR protein quantified immunoreactivity in particular brain regions whereas the current report does not have the same anatomical resolution. Second, our previous reports compared groups of subordinates and breeders without controlling for duration of pairing, number of litters/size of colony, presence of new litter, etc., whereas the breeders herein were all collected exactly 30 days after the birth of their first litter. Finally, patterns of AR protein and mRNA expression are not always complementary (for review, see [34]). This raises the possibility that, in this case, non-coordinate regulation of AR may be achieved via differences in mRNA translation. Although this is an uncommon mechanism, naked mole-rats have a ribosomal organization which is unique in eukaryotic organisms [35] and could very well have similarly unique regulation of protein translation. It will be important to directly quantify protein and mRNA in the same animals to test these alternatives.

In addition to the increase in Ar mRNA after the production of a litter, male naked mole-rats exhibit other neuroendocrine changes similar to new fathers from other species. For example, circulating E2 increased and Pgr mRNA decreased in new naked mole-rat fathers. In humans, an increase in E2 has been reported in men becoming fathers that lasted from the last month of pregnancy to the first month after pregnancy [28]. Reduced Pgr has also been associated with increased paternal behavior in mice: Pgr-knockout mice and pharmacological antagonism of Pgr increased parental care and decreased infanticide behavior without diminishing male-male aggression, whereas increased P has the opposite effect ([36]; for review, see [37]). Alternatively, it is possible that this small decrease in Pgr is not biologically functional. Future experiments could evaluate whether PR has a role similar to other species in male naked mole-rats.

In female naked mole-rats, the first observed endocrine change was an increase in circulating E2 and P 1 month post-removal from their colony. This increase may be indicative of sexual maturation and sexual receptivity. Estrogens are known to trigger, and P to reinforce, the events leading to ovulation in other rodent species (for review, see [38, 39]). E2 is also related to lordosis behavior in rodents, and P acting via progesterone receptor amplifies this behavior (for review, see [40]). Increases in Esr1 and Pgr mRNA were not apparent in female naked mole-rats until after the production of a litter; thus, it is possible that increases in circulating E2 and P are responsible for the increase in diencephalic mRNA expression of the corresponding receptors. However, it is also possible that this increase in gene expression is related to new motherhood, as in other species, increases in Esr1 are also associated with maternal care [41].

The timing of our increase in E2 and P at 1 month post-removal is later then we predicted based on reports of increased P in the urine of female naked mole-rats approximately 1 week after removal from the colony [12, 16]. This might reflect different hormone concentrations in urine versus blood, differences in assay sensitivity, or other methodological differences concerning sample collection. We collected the blood at a single time point whereas previous reports collected urine at multiple times per animal [12, 16]. Alternatively, the variability between experiments might reflect individual differences in the timing of reproductive maturation in subordinate naked mole-rats. New pairs of opposite-sex subordinates show substantial variability in the amount of time to produce their first litter, ranging from 2 months to over 2 years [Holmes, unpublished observations]. The five breeding pairs included here ranged from 3.5 to 6.5 months to produce their first litter. It would be interesting to test whether variability in the onset of endocrine changes corresponds with the variability in the time it takes a new pair of naked mole-rats to produce a litter.

Unexpectedly, we observed a small increase in Cyp19a1 mRNA expression 1 week post-removal in females and a marked increase in females after the production of a litter. These increases were not seen in males and could be related to the dominant status the breeding female undertakes in the colony. Aromatase has been linked to aggression and dominance in other species. For example, Unger et al. [42] report that hypothalamic aromatase regulates aggression in both male and female mice, and Huffman, O’Connell, and Hofmann [43] report that an aromatase inhibitor decreases aggression in cichlid fish [42, 43]. Currently, little is known about how naked mole-rat female breeders maintain their dominance over others in the colony; therefore, it would be of interest to evaluate the role of aromatase given the increase in Cyp19a1 during the transition to breeder status. Another possibility, which is not mutually exclusive, is that increased Cyp19a1 mRNA could be related to female sexual behavior. Observations from Cyp19a1 knockout mice suggest that aromatase may be necessary for female-typical sexual behaviors (for review, see [44]). Similarly in prairie voles, aromatase inhibitors reduce female sexual behaviors [45]. Regardless, dominance and sexual behaviors are both important for a naked mole-rat female to establish and maintain her position at the top of the social hierarchy. Evaluating the role that aromatase plays in social status and reproduction in naked mole-rats will shed light on key features of eusociality in mammals, including how breeding females suppress reproductive maturation in their colony mates. Given that aromatase converts testosterone to estradiol and we report low levels of circulating testosterone in all female groups, the increased Cyp19a1 mRNA in breeding females could be associated with aromatase acting on other androgenic substrates. Alternatively, our single sampling paradigm might have missed cyclic variation in circulating androgens or reflect a dissociation between androgens in the blood and brain.

Finally, the group differences we see in body weight are consistent with previous reports that breeders, particularly queens, are often the largest animals in the colony (e.g., [9]). Furthermore, queen removal results in increased body weight of both male and female subordinates [14], suggesting the release from reproductive suppression triggers growth in non-reproductive animals. Indeed, new queens demonstrate remarkable vertebral lengthening associated with parity [46], which might also be associated with the release from reproductive suppression and representative of a pubertal growth spurt [47].

The present study indicates that naked mole-rats exhibit sex-specific neuroendocrine changes when transitioning from subordination to reproductive maturation and social dominance. Naked mole-rats transitioning to breeding status resemble other mammalian species during puberty and begin to exhibit endocrine changes as early as 1 week following removal from the reproductive suppression of their colony. The sex-specific activation of gonadal steroid hormones and their receptors in the brain are clearly directly linked to reproduction. However, some of the neuroendocrine changes we report might also reflect sex-specific demands in colony living (e.g., dominance in breeding females and paternal behavior in breeding males). Given that reproductive hierarchies and large social colonies are two defining features of eusociality, these data reveal much about how this extreme form of social organization occurs in mammals. In conclusion, these data (a) confirm that naked mole-rats show sex differences in circulating gonadal hormones only when released from reproductive suppression, (b) demonstrate that latent sex differences in gonadal steroid hormone receptors exist in the brain of these remarkably monomorphic animals, and (c) provide future directions for evaluating the neuroendocrine basis of eusociality in naked mole-rats.