Developmental GnRH Signaling Is Not Required for Sexual Differentiation of Kisspeptin Neurons but Is Needed for Maximal Kiss1 Gene Expression in Adult Females

Abstract

Kisspeptin, encoded by Kiss1, stimulates reproduction. In rodents, one Kiss1 population resides in the hypothalamic anterior ventral periventricular nucleus and neighboring rostral periventricular nucleus (AVPV/PeN). AVPV/PeN Kiss1 neurons are sexually dimorphic (greater in females), yet the mechanisms regulating their development and sexual differentiation remain poorly understood. Neonatal estradiol (E₂) normally defeminizes AVPV/PeN kisspeptin neurons, but emerging evidence suggests that developmental E₂ may also influence feminization of kisspeptin, although exactly when in development this process occurs is unknown. In addition, the obligatory role of GnRH signaling in governing sexual differentiation of Kiss1 or other sexually dimorphic traits remains untested. Here, we assessed whether AVPV/PeN Kiss1 expression is permanently impaired in adult hpg (no GnRH or E₂) or C57BL6 mice under different E₂ removal or replacement paradigms. We determined that 1) despite lacking GnRH signaling in development, marked sexual differentiation of Kiss1 still occurs in hpg mice; 2) adult hpg females, who lack lifetime GnRH and E₂ exposure, have reduced AVPV/PeN Kiss1 expression compared to wild-type females, even after chronic adulthood E₂ treatment; 3) E₂ exposure to hpg females during the pubertal period does not rescue their submaximal adult Kiss1 levels; and 4) in C57BL6 females, removal of ovarian E₂ before the pubertal or juvenile periods does not impair feminization and maximal adult AVPV/PeN Kiss1 expression nor the ability to generate LH surges, indicating that puberty is not a critical period for Kiss1 development. Thus, sexual differentiation still occurs without GnRH, but GnRH or downstream E₂ signaling is needed sometime before juvenile development for complete feminization and maximal Kiss1 expression in adult females.

The neuropeptide kisspeptin, encoded by Kiss1, is critical for puberty and fertility (1, 2). In rodents, kisspeptin-synthesizing neurons reside in the hypothalamic arcuate nucleus and in the hypothalamic continuum encompassing the anterior ventral periventricular nucleus and neighboring rostral periventricular nucleus (AVPV/PeN) (3–5). In adulthood, estradiol (E₂) suppresses Kiss1 levels in the arcuate but robustly elevates Kiss1 expression in the AVPV/PeN (5–7). Because kisspeptin directly stimulates GnRH activation (8–10), and AVPV/PeN Kiss1 neurons express estrogen receptor α (ERα) (7), these neurons likely mediate E₂'s positive feedback effects on GnRH/LH secretion (ie, the female LH surge) (11). Supporting this, in female rodents, there is an evening increase in AVPV/PeN Kiss1 neuronal activation that coincides with the circadian onset of GnRH neuron activation and the LH surge (12–14). Moreover, Kiss1r knockout (KO) female mice, which lack kisspeptin-Kiss1r signaling, cannot produce an E₂-induced LH surge (15, 16).

Supporting a role in the sexually dimorphic LH surge event, AVPV/PeN kisspeptin neurons are themselves sexually dimorphic in cell number and Kiss1 mRNA expression (greater in females than males) (3, 5, 17). Although detectable Kiss1 and kisspeptin expression in the rodent AVPV/PeN does not first emerge until juvenile life (3, 18–20), we and others have demonstrated that sexual differentiation of AVPV/PeN Kiss1 neurons is permanently induced via developmental sex steroid signaling in the neonatal period (ie, an organizational effect of neonatal steroids) (5, 17, 21). Neonatal female rodents given testosterone or E₂ at or shortly after birth have reduced, male-like numbers of AVPV/PeN Kiss1 neurons later in adulthood (5, 17, 22), regardless of adult sex steroid levels. Conversely, removing sex steroids via castration from neonatal males has the opposite effect, producing higher female-like Kiss1 expression in the AVPV/PeN in adulthood (17). The precise mechanisms by which perinatal E₂ directs the development and sexual differentiation of AVPV/PeN Kiss1 neurons are still unknown, but may include epigenetic modifications of Kiss1 expression (23).

Organizational sex differences can only be induced during specific developmental ages or critical periods, outside of which sexual differentiation cannot be altered. For example, sex steroid treatment at birth permanently masculinizes the POA volume in rats, whereas testosterone given later in juvenile life no longer masculinizes this region (24, 25). Although the presence of a neonatal critical period is well-supported, it is currently unresolved whether puberty also represents another critical period for kisspeptin sexual differentiation. Several studies, primarily in hamsters, suggest that the pubertal phase represents a critical period for some sexually dimorphic traits (26–29). Moreover, Clarkson et al (30) observed that gonadectomy in female mice before puberty (on postnatal day [PND] 15) reduced AVPV/PeN kisspeptin levels in early adulthood, and that E₂ replacement initiated during puberty could rescue this deficit. It was therefore proposed that the developmental trajectory of kisspeptin neurons in the female AVPV/PeN is further feminized by E₂ acting in a pubertal critical period. However, the Clarkson data are difficult to interpret because group differences in circulating E₂ at the time of brain collection were not controlled for (E₂ transiently up-regulates adult kisspeptin via activational effects [7]). Therefore, contributions of potential organizing effects of E₂ during a critical pubertal period versus activational effects of E₂ at the time of killing remain unsorted. Regardless, additional data from female aromatase KO mice and hypogonadal (hpg) mice (lacking GnRH and hence, gonadal E₂) suggest that developmental E₂ exposure may indeed influence maximal female-like kisspeptin development. Both aromatase KO and hpg adult females display reduced kisspeptin protein staining in the AVPV/PeN (assessed via immunohistochemistry), even after short-term E₂ replacement in adulthood (31, 32). These findings suggest that E₂ signaling is required at some point in development or adulthood for display of maximal female-like kisspeptin expression (ie, complete feminization). However, exactly when E₂ is needed to complete kisspeptin feminization is unknown. Furthermore, because similarly aged hpg males and females were not directly compared (31), it remains unclear if AVPV/PeN kisspeptin is still sexually dimorphic in hpg mice or, alternatively, if kisspeptin levels in hpg females are lowered all the way to male levels. Indeed, the role—if any—of GnRH signaling in directly or indirectly governing sexual differentiation remains a contentious issue. Intriguingly, hpg male mice were recently reported to display a normal neonatal testosterone surge at birth (33), suggesting that sexual differentiation might still occur despite absent GnRH; however, this was not assessed.

We sought to answer several unresolved issues regarding AVPV/PeN Kiss1 sexual differentiation and development. We determined 1) if sexual differentiation of AVPV/PeN Kiss1 is impaired in hpg mice (no GnRH signaling), such that adult hpg males and females have similar Kiss1 levels; 2) if AVPV/PeN Kiss1 mRNA expression is, like previously reported kisspeptin protein levels, reduced in adult hpg females, thereby distinguishing whether the lower kisspeptin protein in hpg mice reflects impaired Kiss1 gene expression or impaired translation; 3) if short-term or long-term E₂ exposure in adulthood can rescue reduced Kiss1 levels in hpg females; 4) if pubertal E₂ exposure can permanently restore maximal Kiss1 expression in adult hpg females; and 5) if removal of gonadal E₂ during the pubertal or juvenile periods in C57BL6 females permanently reduces AVPV/PeN Kiss1 levels or LH surges in adulthood.

Materials and Methods

Animals

C57BL6 mice or hypogonadal (hpg) and wild-type (WT) littermates (Jackson Labs, Bar Harbor, Maine) were used. Adult hpg homozygous mice lack GnRH and are therefore infertile and have undeveloped gonads and absent gonadal sex steroids. Hpg litters were genotyped after weaning (PND 21) via PCR analysis of tail DNA. In all experiments, weaned mice were provided food and water ad libitum and housed 2 to 3 under a 12:12 light-dark cycle (lights off at 1800 h). All experiments were approved by the local University Animal Care and Use Committee.

Gonadectomies and hormone treatments

Either in adulthood or in development, mice were anesthetized with isoflurane and bilaterally gonadectomized (GDX) through a single ventral midline incision. The abdominal musculature was sutured with sterile chromic gut and the skin incision closed with sterile wound clips (adults) or sterile suture (juvenile/prepubertal mice).

Because E₂ transiently up-regulates AVPV/PeN Kiss1 expression in adults via activational effects (7, 23), E₂ levels were equalized among groups before brain collection. Some adult GDX mice (see specific experiments) received a subcutaneous Silastic implant filled with 4 mm of 17β-E₂ diluted 1:5 with cholesterol. These E₂ implants produce high physiological blood E₂ levels and significantly up-regulate Kiss1 expression in the AVPV/PeN of adult mice (7, 34). In experiment 4, peripubertal female mice were GDX and temporarily given a smaller 2 mm Silastic implant containing E₂ diluted 1:20 with cholesterol, which a pilot experiment determined produced E₂ levels in the adult proestrus range. These peripubertal E₂ implants were removed 8 days later, and the skin incision was closed with a sterile wound clip.

Experiment 6 assessed the ability of females to generate an E₂-induced LH surge, using a validated surge paradigm in which adult GDX females were given a subcutaneous implant containing 0.65 μg 17β-E₂ dissolved in sesame oil, as previously described (13, 16). This E₂ treatment produces proestrus-like E₂ levels and induces robust evening LH surges 2 days later (13, 16, 35).

Brain collection and in situ hybridization (ISH)

Mice were anesthetized with isoflurane and brains were collected and frozen on dry ice before storage at −80°C. Frozen brains were sectioned on a cryostat into 5 sets of 20-μm sections, thaw-mounted on Superfrost-plus slides, and stored at −80°C until assaying. Single-label ISH for Kiss1 mRNA was performed using a validated murine Kiss1 riboprobe (4). Briefly, one set of slide-mounted sections encompassing the entire AVPV/PeN region was fixed in 4% paraformaldehyde, pretreated with acetic anhydride, rinsed in 2× SSC (sodium citrate, sodium chloride), delipidated in chloroform, dehydrated in ethanols, and air-dried. Radiolabeled (³³P) antisense Kiss1 riboprobe (0.04 pmol/mL) was combined with 1/20 vol yeast tRNA, heat-denatured, added to hybridization buffer, and applied to each slide (100 μL/slide). Slides were coverslipped and placed in a humidity chamber at 55°C for 18 hours. Following hybridization, slides were washed in 4× SSC and then placed into RNAse (37 mg/mL RNAse A in 0.15M sodium chloride, 10 mM Tris, 1 mM EDTA, pH 8.0) for 30 minutes at 37°C, then in RNAse buffer without RNase at 37°C for 30 minutes. After washing in 2× SSC at room temperature, slides were washed in 0.1× SSC at 62°C, dehydrated in ethanols, and air-dried. Slides were dipped in Kodak NTB emulsion, air-dried, and stored at 4°C for 4 to 6 days (depending on the assay) before being developed and coverslipped.

For double-label ISH of ERα in Kiss1 neurons (experiment 3), the single-label protocol was used with slight modification. Briefly, radiolabeled (³³P) antisense ERα (0.04 pmol/mL) and digoxigenin (DIG)-labeled Kiss1 (1:500) riboprobes were combined with tRNA, denatured, dissolved together in hybridization buffer, and applied to each slide (100 μL/slide) before overnight hybridization at 55°C. After day 2's 62°C washes, slides were incubated in 2× SSC with 0.05% Triton X-100 containing 3% normal sheep serum for 1 hour at room temperature and then incubated overnight with anti-DIG antibody conjugated to alkaline phosphatase (Roche, Indianapolis, Indiana; diluted 1:500 in buffer 1 containing 1% normal sheep serum and 0.3% Triton X-100). Slides were then washed with buffer 1 and incubated with Vector Red alkaline phosphatase substrate (Vector Labs, Philadelphia, Pennsylvania) for 1 hour at room temperature. Slides were then air-dried, dipped in emulsion, stored at 4°C, and developed 7 days later.

Quantification of ISH data and statistical analyses

ISH slides were analyzed blind to treatment using an automated custom grain-counting software (Dr Don Clifton, University of Washington, Seattle, Washington) that counts the number of silver grain clusters representing cells, as well as the number of silver grains in each cell (a semiquantitative index of mRNA content per cell) (36). Cells were considered Kiss1 positive when the number of silver grains in a cluster exceeded that of background by 3-fold. A relative measure of total mRNA in the AVPV/PeN region was calculated by multiplying the total number of cells by the number of silver grains per cell. For double-label ISH, red fluorescent DIG (Kiss1) cells were identified under fluorescence microscopy and the grain-counting software was used to quantify the number of silver grains (ERα mRNA) overlying each cell. Signal-to-background ratios for individual cells were calculated by the program, and a cell was considered double-labeled if its ratio was >3. All data are expressed as the mean ± SEM for each group. Differences were analyzed by ANOVA, followed by post-hoc comparisons via Fisher (protected) least significant difference test. Statistical significance was set at P < .05.

Experiment 1: Is Kiss1 expression in the AVPV/PeN still sexually differentiated in hpg mice?

We recently reported that, despite having no GnRH signaling (and hence, no downstream gonadotropin signaling), neonatal hpg male mice still display normal testosterone secretion at birth (33). Because neonatal testosterone secretion drives sexual differentiation, our prior finding suggested that sexual differentiation may occur normally in hpg mice, but this was not studied. This experiment directly tested whether AVPV/PeN Kiss1 sex differences were still present in adult hpg mice. WT and hpg males and females were GDX on PND 38 and subcutaneously implanted with E₂ to control for activational effects of sex steroids. One week later (PND 45), mice were killed and their brains collected (n = 5–7/group). Single-label ISH for Kiss1 in the AVPV/PeN was performed and the number of Kiss1-expressing cells, the relative level of Kiss1 mRNA per cell, and the relative amount of total Kiss1 mRNA were measured.

Experiment 2: Does long-term E₂ exposure in adulthood rescue the submaximal Kiss1 expression in hpg females?

Experiment 1 showed that, despite still being sexually differentiated, adult hpg females have fewer detectable Kiss1 neurons relative to WT females, even after short-term (1 wk) E₂ exposure in adulthood. However, because hpg females have never been exposed to E₂ during the juvenile, pubertal, or early adult periods, it may be that hpg females merely require a longer E₂ exposure than just 1 week to complete the development and feminization of their Kiss1 system. This experiment therefore tested if a chronic, longer term exposure to E₂ could rescue the impaired Kiss1 expression in adult hpg females and restore their Kiss1 levels to those of WT females. First, to confirm our experiment 1 finding and also assess whether Kiss1 is in fact E₂-responsive in hpg mice (contrary to a previous report looking at kisspeptin-ir levels [31]), we examined adult hpg and WT females that were GDX on PND 38 and given either an E₂ implant (n = 7–8/group) or nothing (n = 4–7/group); brains from all mice were collected 1 week later (PND 45) and analyzed for Kiss1 in the AVPV/PeN using single-label ISH, as in experiment 1. Next, a separate cohort of hpg and WT females were ovariectomized (OVX) on PND 38 and chronically implanted with E₂ (n = 8–9/genotype); a control group of PND 38 WT females was similarly GDX but not given any E₂ implant (n = 5). All mice were then killed 22 days later (PND 60) and their brains collected. To prevent exhaustion of the E₂ implants, they were removed after 11 days and replaced with fresh implants for the remaining 11 days. Brains were analyzed for Kiss1 mRNA levels in the AVPV/PeN using ISH.

Experiment 3: Do AVPV/PeN Kiss1 neurons of hpg females express less ERα than WT females?

Experiment 2 determined that although the Kiss1 gene is E₂-responsive, AVPV/PeN Kiss1 levels in hpg females remain submaximal even after chronic long-term E₂ exposure in adulthood. We hypothesized that this reduced Kiss1 levels in hpg females may reflect a lower degree of ERα coexpression in Kiss1 neurons of hpg than WT females, thereby producing lower Kiss1 up-regulation in the former genotype. To test this, we performed double-label ISH to assess and quantify the degree of ERα expression in AVPV/PeN Kiss1 neurons in adult hpg and WT females (n = 4/genotype). The percentage of AVPV/PeN Kiss1 neurons expressing ERα and the relative amount of ERα mRNA per Kiss1 cell were determined for each genotype.

Experiment 4: Does E₂ exposure during the pubertal period rescue the incomplete feminization of Kiss1 expression in hpg females?

Experiments 2 and 3 determined that although the Kiss1 gene is E₂-responsive and expresses normal levels of ERα, Kiss1 levels in hpg females remain submaximal even after chronic long-term E₂ exposure in adulthood. This suggests a developmental problem rather than lack of adult sex steroid exposure. Previous rodent studies suggested that the pubertal period is a critical period for sexual differentiation of some brain traits, and this may possibly also be true for AVPV/PeN kisspeptin. This experiment tested whether the lower Kiss1 levels in adult hpg females reflects an absence of developmental E₂ exposure during the pubertal period. In our mouse colony, female puberty (reflected by vaginal opening) typically occurs between PND 26 and PND 29. WT and hpg females were therefore GDX just before puberty, on PND 22, and simultaneously implanted with E₂. The E₂ implant was left in for 8 days to span the pubertal period and then removed, after which all mice aged without additional E₂ exposure. At PND 56, females of both genotypes were implanted with either E₂ (to control for activational effects of adult E₂) or nothing and brains were collected 1 week later (PND 63) (n = 6–8/genotype). Single-label ISH for Kiss1 in the AVPV/PeN was performed as in prior experiments.

Experiment 5: Is E₂ required during the juvenile and/or pubertal period for complete feminization of AVPV/PeN Kiss1 neurons?

Our previous experiments suggested that E₂ is needed sometime in development for complete feminization of the AVPV/PeN Kiss1 system. Previous rodent studies suggested that puberty is a critical period for sexual differentiation of some traits, but this has not been definitively studied for AVPV/PeN kisspeptin. Here, we tested this possibility in normal C57BL6 females by removing gonadal E₂ before puberty and assessing the impact—if any—on Kiss1 levels later in adulthood. We also assessed the necessity of E₂ exposure at even younger ages before puberty by testing whether permanent removal of gonadal E₂ before AVPV/PeN Kiss1 expression first initiated in juvenile development disrupts feminization of this Kiss1 system. C57BL6 females were GDX on either PND 14 (before the prepubertal and pubertal periods) or PND 9 (before the juvenile period and the first detectable AVPV/PeN Kiss1 expression on PND 10 [18]). All GDX females were then aged to PND 35, when they were implanted with either E₂ (n = 8–9) or nothing (n = 5) and killed 1 week later (PND 42). Control C57BL6 females were left gonadally intact throughout all of development, GDX, and implanted with E₂ (n = 5–6) or nothing (n = 5–7) on PND 35, and similarly killed on PND 42. Brains were collected and Kiss1 analyzed via ISH.

Because sexually dimorphic AVPV/PeN kisspeptin is thought to drive the circadian-timed LH surge in females, a complementary experiment examined whether females lacking developmental E₂ could generate LH surges in adulthood. C57BL6 females were GDX on PND10 and aged in the absence of juvenile and pubertal E₂ exposure. On PND 54, they were implanted with an E₂ paradigm that normally induces evening LH surges 2 days later. Control females remained gonadally intact throughout development, were GDX on PND 48, and similarly implanted with E₂ on PND 54. All mice were killed on PND 56, in either the am (when surges do not occur) or the pm (at lights off, when surges normally occur) (n = 5–6/group). Blood was collected and serum LH levels were determined using RIA (University of Virginia Ligand Assay Laboratory, Charlottesville, Virginia).

Results

Experiment 1: Kiss1 expression in the AVPV/PeN is sexually dimorphic in hpg mice

This experiment tested if AVPV/PeN Kiss1 expression is still sexually dimorphic or not in hpg mice (lacking GnRH signaling) with equalized adult E₂ levels. As expected, adult WT females had significantly higher Kiss1 expression than WT males (P < .01 for Kiss1 cell number, Kiss1 mRNA per cell, and total Kiss1 mRNA levels; Figures 1 and 2). Interestingly, despite the lifetime absence of GnRH signaling, AVPV/PeN Kiss1 was still sexually dimorphic in adult hpg mice, with hpg females having significantly higher Kiss1 cell number (by 33%), Kiss1 mRNA/cell (by 41%), and total Kiss1 mRNA (by 87%) than hpg males (P < .01 for each measure; Figures 1 and 2). However, feminization of Kiss1 appeared submaximal in hpg mice: Kiss1 cell number and total Kiss1 mRNA levels were both approximately 40% lower in hpg females than WT females (P < .01; Figure 2). Likewise, defeminization (or masculinization) of Kiss1 expression in hpg males was not maximal: for all measures, Kiss1 levels were significantly higher (P < .05) in hpg males than WT males (25% to 86% higher, depending on the measure; Figure 2). Thus, while Kiss1 expression is still sexually differentiated in hpg mice, there may be a developmental or adulthood requirement for GnRH or downstream E₂ signaling for maximal Kiss1 expression (perhaps denoting complete feminization) in females.

Figure 1.

Representative photomicrographs of Kiss1 gene expression, assessed via in situ hybridization, in the AVPV/PeN of adult hpg and WT male and female mice. All mice were treated with E₂ for 1 week before brain collection. 3V, third ventricle.

Figure 2.

Sexually dimorphic Kiss1 expression in the AVPV/PeN of E₂-treated adult (d45) hpg and WT males and females. (A) Mean number of Kiss1 cells in the AVPV/PeN. (B) Mean relative levels of Kiss1 mRNA per cell in the AVPV/PeN. (C) Mean relative levels of total Kiss1 mRNA in the entire AVPV/PeN region. For all measures, Kiss1 was sexually dimorphic in hpg mice (higher in females). Bars labeled with different letters are significantly different from each other (P < .05).

Experiment 2: AVPV/PeN Kiss1 expression is E₂-responsive in hpg females but still lower than in WTs, even after chronic adult E₂ exposure

Adult hpg and WT females were OVX on PND 38, treated for 1 week with E₂ or vehicle, and their brains were analyzed for AVPV/PeN Kiss1 expression. As expected, WT OVX females had low Kiss1 expression, whereas WT OVX+E₂ females had significantly higher Kiss1 expression (Kiss1 cell number, Kiss1 mRNA per cell, and total Kiss1 mRNA levels; P < .01 for each measure; Supplemental Figures 1 and 2, which are published on The Endocrine Society's Journals Online web site at http://endo.endojournals.org). Likewise, OVX+E₂ hpg females had significantly higher Kiss1 cell number, Kiss1 mRNA/cell, and total Kiss1 mRNA than OVX hpg females without E₂ (P < .01 for each measure; Supplemental Figures 1 and 2), indicating the Kiss1 gene is still E₂ responsive. However, as in experiment 1, Kiss1 levels were significantly lower in E₂-treated hpg females than E₂-treated WTs (P < .05; Supplemental Figure 2). Thus, while Kiss1 expression is still responsive to E₂ in hpg females, the maximal level of Kiss1 expression attained with short-term (1 wk) E₂ exposure is significantly lower (by 38%-45%, depending on the measure) in hpg than WT mice. We therefore next tested if longer term exposure to E₂ in adulthood could rescue the submaximal Kiss1 levels in hpg females. Adult WT females given chronic E₂ (3 wk) displayed the expected elevation in AVPV/PeN Kiss1 levels relative to WT OVX without E₂ (P < .01; Figure 3). Likewise, adult hpg females given long-term E₂ for 3 weeks had robustly elevated Kiss1 levels, significantly greater than OVX females (P < .01; Figure 3). However, Kiss1 levels in chronically E₂-treated hpg females were still significantly lower than in E₂-treated WT females, particularly for Kiss1 cell number (25% lower) and total Kiss1 mRNA levels (33% lower) (P < .01 for each, Figure 3). Kiss1 mRNA per cell also showed a nonsignificant trend for lower levels in hpg females (P = .10).

Figure 3.

Kiss1 expression in the AVPV/PeN of adult (d60) hpg and WT females with or without 22 days of chronic adulthood E₂ exposure. (A) Mean number of Kiss1 cells in the AVPV/PeN. (B) Mean relative levels of Kiss1 mRNA per cell in the AVPV/PeN. (C) Mean relative levels of total Kiss1 mRNA in the entire AVPV/PeN region. Bars labeled with different letters are significantly different from each other (P < .05).

Experiment 3: AVPV/PeN Kiss1 neurons of hpg females express normal ERα levels

Experiment 2 determined that AVPV/PeN Kiss1 levels in hpg females remain submaximal even after long-term E₂ exposure in adulthood. Here, we determined whether the reduced Kiss1 expression in hpg females reflects lower ERα expression in their Kiss1 neurons. We found that ERα was highly expressed in the AVPV/PeN of both hpg and WT females and present in most Kiss1 cells in both genotypes (Figure 4A). Quantitatively, neither the percentage of AVPV/PeN Kiss1 neurons expressing ERα nor the relative amount of ERα mRNA per Kiss1 cell differed significantly between genotypes (Figure 4, B and C).

Figure 4.

ERα-Kiss1 coexpression in the AVPV/PeN of adult hpg and WT females. (A) Representative photomicrographs of ERα expression in AVPV/PeN Kiss1 neurons, assessed via double-label in situ hybridization, in hpg and WT females. Yellow arrows designate examples of colabeled cells. Blue arrows designate example Kiss1 cells lacking ERα. 3V, third ventricle. (B) Mean number of Kiss1 cells in the AVPV/PeN that coexpress ERα. (C) Mean relative levels of ERα mRNA (silver grains) per Kiss1 cell in the AVPV/PeN. There were no genotype differences in any measure.

Experiment 4: Pubertal E₂ exposure is not sufficient to rescue the submaximal Kiss1 levels in hpg females

This experiment tested whether the lower Kiss1 levels observed in adult hpg females reflects an absence of organizing E₂ exposure during the pubertal period, a potential critical period. WT and hpg females were OVX on PND 22 and given E₂ for 8 days and then left to age without additional E₂ exposure until PND 56. WT females given E₂ in adulthood had significantly higher Kiss1 expression in all measures than WT females not given E₂ in adulthood (P < .05; Figure 5). A similar pattern was observed for hpg females with and without adulthood E₂ (P < .05; Figure 5). However, as in previous experiments, hpg females still had significantly lower Kiss1 levels than WT females of the same treatment (P < .05; Figure 5), even though all mice were exposed to E₂ during the pubertal period. Thus, pubertal E₂ exposure is not sufficient to promote complete feminization of AVPV/PeN Kiss1 neurons.

Figure 5.

Kiss1 expression in the AVPV/PeN of adult (d63) hpg and WT females after pubertal E₂ treatment from PND 22 to PND 30. (A) Mean number of Kiss1 cells in the AVPV/PeN. (B) Mean relative levels of Kiss1 mRNA per cell in the AVPV/PeN. (C) Mean relative levels of total Kiss1 mRNA in the entire AVPV/PeN region. Bars labeled with different letters are significantly different from each other (P < .05).

Experiment 5: E₂ is not required during the pubertal or juvenile periods for complete feminization of AVPV/PeN Kiss1 neurons

Here, we first tested in C57BL6 mice whether permanently removing E₂ exposure before the prepubertal and pubertal periods impacts the development and feminization of Kiss1 in the AVPV/PeN. Control females left gonadally intact throughout development and then OVX+E₂ on PND 35 displayed the expected high levels of Kiss1 one week later compared to control OVX females not given E₂ (P < .01; Figure 6). Similarly, females that were prepubertally OVX on PND14 and allowed to age without pubertal E₂ exposure demonstrated maximal Kiss1 levels after E₂ treatment administered on PND 35 (Figure 6). There was no difference in any Kiss1 measure between E₂-treated females OVX before the prepubertal period (on PND 14) and control E₂-treated females OVX on PND 35 (Figure 6), indicating that E₂ is not required during puberty for feminization and maximal Kiss1 expression. Likewise, in the complementary experiment, C57BL6 females that were OVX on PND 9, before AVPV/PeN Kiss1 expression first emerges in juvenile development, demonstrated maximal Kiss1 levels after E₂ treatment later in young adulthood (given for 1 wk on PND 35; Figure 7). There was no significant difference in any Kiss1 measure between females OVX before the juvenile period (on PND 9) or control females OVX on PND 35 (Figure 7), indicating that gonadal E₂ is not required during the juvenile period for complete feminization and maximal AVPV/PeN Kiss1 expression.

Figure 6.

Kiss1 expression in the AVPV/PeN of adult (d42) C57BL6 females that were OVX either after puberty, on PND 35, or before puberty, on PND 14. On PND 35, mice from each OVX group were given either an E₂ implant for 1 week or no implant. (A) Mean number of Kiss1 cells in the AVPV/PeN. (B) Mean relative levels of Kiss1 mRNA per cell in the AVPV/PeN. (C) Mean relative levels of total Kiss1 mRNA in the entire AVPV/PeN region. Bars labeled with different letters are significantly different from each other (P < .05).

Figure 7.

Kiss1 expression in the AVPV/PeN of adult (d42) C57BL6 females that were OVX either after puberty, on PND 35, or before the juvenile period and time of first Kiss1 expression, on PND 9. On PND 35, mice from each OVX group were given either an E₂ implant for 1 week or no implant. (A) Mean number of Kiss1 cells in the AVPV/PeN. (B) Mean relative levels of Kiss1 mRNA per cell in the AVPV/PeN. (C) Mean relative levels of total Kiss1 mRNA in the entire AVPV/PeN region. Bars labeled with different letters are significantly different from each other (P < .05).

Because sexually dimorphic AVPV/PeN kisspeptin neurons govern the sexually dimorphic LH surge, we also examined whether adult females lacking developmental E₂ could still generate normal LH surges. C57BL6 females OVX on PND 10 and given E₂ later on PND 54 displayed robust evening LH surges, similar to control females that were left gonadally intact throughout development and OVX on PND 48 (Figure 8). LH levels of both pm groups were significantly higher than am control females (P < .05; Figure 8). Thus, E₂ is not required after PND 10 for complete sexual differentiation of the LH surge mechanism.

Figure 8.

LH surge in mice with and without juvenile and pubertal E₂ exposure. C57BL6 females were OVX either before the juvenile period, on PND 10, or later in adulthood, on PND 48. All mice were given a positive feedback dosing of E₂ on PND 54 and LH was measured 2 days later in either the am or pm. Bars labeled with different letters are significantly different from each other (P < .05).

Discussion

Sexual differentiation of AVPV/PeN Kiss1 neurons may underlie several sexually dimorphic processes, including the E₂-induced LH surge (positive feedback) and perhaps puberty. Although perinatal sex steroids, primarily E₂, permanently establish the sex difference in AVPV/PeN Kiss1 expression, exactly how and when E₂ governs this process remains poorly understood. Here, we provide new evidence regarding the timing and necessity of GnRH and E₂ signaling in the sexual differentiation and development of AVPV/PeN Kiss1 expression. We demonstrate that adult hpg females, who permanently lack GnRH (and hence, E₂), still have markedly higher AVPV/PeN Kiss1 levels than hpg males, indicating that sexual differentiation does not require GnRH signaling. However, Kiss1 gene expression in adult hpg females is markedly reduced compared to WT females, and these submaximal Kiss1 levels cannot be rescued by short- or long-term adulthood E₂ treatment. This determination, in combination with our finding of normal Kiss1-ERα coexpression in adult hpg females, suggests a developmental rather than adulthood defect. We show that this developmental deficit is unlikely to be during the pubertal period, as hpg females exposed to E₂ during the pubertal stage still exhibit submaximal Kiss1 levels in adulthood. Thus, puberty is not a critical period for Kiss1 sexual differentiation, a finding confirmed by normal (complete) feminization of Kiss1 in C57BL6 females after ovary removal on PND 9 or PND 14. Overall, our results indicate that while GnRH signaling is not needed for sexual differentiation of Kiss1, GnRH (likely via downstream E₂ secretion) is still needed sometime before juvenile development in females for complete feminization and maximal expression of AVPV/PeN Kiss1 neurons.

Sexual differentiation of neural circuits, including AVPV/PeN Kiss1, is induced primarily by gonadal sex steroid secretion in perinatal males but not females. Because adult sex steroid secretion is dependent on upstream GnRH secretion, there is an assumption that perinatal androgen secretion (and hence, sexual differentiation) is also governed by GnRH signaling, but this has not been directly assessed. Our present results unquestionably demonstrate that AVPV/PeN Kiss1 is still sexually dimorphic in the absence of GnRH signaling: adult hpg females had ∼33% more Kiss1-expressing cells and ∼90% more Kiss1 mRNA than hpg males. Therefore, GnRH signaling during the perinatal period (or any other developmental stage) is not required for sexual differentiation, at least for Kiss1. This conclusion is supported by our recent finding (33) that newborn hpg males, like WT males, produce a normal neonatal androgen surge, indicating that this androgen secretion—and by extension, sexual differentiation (as confirmed here)—is GnRH-independent.

Although sexual differentiation of AVPV/PeN Kiss1 was clearly present in hpg mice, Kiss1 mRNA expression in hpg females did not reach maximal WT female levels, perhaps reflecting incomplete feminization. Specifically, Kiss1 cell number and total Kiss1 mRNA were 38% to 45% lower in hpg than WT females under controlled E₂ milieus. This impairment in female hpg Kiss1 mRNA expression confirms and extends 2 recent reports of reduced AVPV/PeN kisspeptin protein in adult hpg and aromatase KO females (31, 32). Given the strong resemblance of the submaximal Kiss1 phenotype in hpg females to that of aromatase KO females, the lower AVPV/PeN Kiss1 levels in hpg females could reflect impaired organizational and/or activational effects of E₂ (owing to E₂'s dependence on upstream GnRH signaling after the neonatal period). However, because neither short-term (1 wk) nor long-term (3 wk) E₂ treatment in adulthood fully restored hpg females' Kiss1 levels to WT female levels (still ∼33% lower in hpg females), we conclude that the submaximal Kiss1 expression in hpg females is not simply due to deficient activational E₂ signaling in adulthood nor is the reduced Kiss1 phenotype due to diminished ERα-Kiss1 coexpression in hpg mice, as demonstrated in experiment 3. We therefore addressed whether maximal female Kiss1 levels might rely on organizational E₂ signaling in pubertal development because a pubertal critical period for other traits exists in hamsters (26–29). In addition, because E₂ exposure earlier in the first week of postnatal life defeminizes (or masculinizes) Kiss1, any feminizing effect of E₂ on Kiss1 would not be at perinatal development. We therefore first tried to rescue the submaximal Kiss1 phenotype by exposing hpg females to E₂ during the pubertal period (PND 22–30). However, such pubertal E₂ treatment was not sufficient to restore hpg females' Kiss1 levels to WT female levels, indicating that the pubertal period is unlikely to be a critical period for Kiss1 sexual differentiation. This conclusion was supported by our findings that C57BL6 females ovariectomized before the prepubertal (PND 14) or juvenile (PND 9) periods and aged thereafter in the absence of gonadal E₂ were able to display maximal AVPV/PeN Kiss1 levels later in adulthood. Thus, E₂ exposure during the juvenile or pubertal periods is not necessary for complete Kiss1 feminization and maximal female Kiss1 expression.

As discussed above, neither the juvenile nor pubertal periods appear to be critical periods for E₂'s ability to organize Kiss1 feminization and enable maximal Kiss1 expression in adulthood. Still, our findings of submaximal Kiss1 expression in hpg females indicate that GnRH signaling, likely via downstream E₂ signaling, is required at some point in development for promoting maximal feminization of Kiss1 neurons. Exactly when E₂ acts for such a process remains unknown, but our data suggest it is before PND 9. However, because E₂ exposure before birth (37) or soon after birth (5, 17) defeminizes (or masculinizes), rather than feminizes, Kiss1 neurons, E₂ would seemingly have to act after this perinatal critical period to induce feminizing effects. Thus, E₂ might act between the perinatal critical period and PND 9 to feminize kisspeptin neurons completely, although this would be a very small temporal window and the ovaries are virtually quiescent in steroid production at this period. Unfortunately, ovariectomizing mice before PND 9 is technically challenging, making it difficult to study via this method. An alternate possibility is that the developmental effects of E₂ are dose-dependent rather than age-dependent, with high neonatal E₂ permanently defeminizing/masculinizing Kiss1 and lower (but not absent) neonatal E₂ enabling complete feminization. Again, this possibility is also difficult to assess, given that mouse neonatal E₂ levels are far too low to measure using conventional hormone assays. Last, brain-derived, locally produced E₂ may possibly contribute to the feminization of kisspeptin neurons, although exactly how GnRH signaling (or its absence in hpg mice) would influence neural E₂ synthesis is unclear.

In conclusion, we show that AVPV/PeN Kiss1 is still markedly sexually dimorphic in hpg mice, indicating that sexual differentiation occurs in the absence of GnRH signaling. However, despite normal ERa-Kiss1 coexpression, Kiss1 expression levels in hpg females are not maximal, even after chronic adulthood E₂ treatment, suggesting an organizational (developmental) rather than activational impairment. Because E₂ exposure to hpg females during the pubertal period does not rescue their lower Kiss1 levels, puberty does not appear to be a critical period for Kiss1 sexual differentiation in mice. Supporting this, gonadal E₂ removal before the pubertal or juvenile periods does not disrupt or prevent maximal Kiss1 expression in adulthood. Overall, these findings demonstrate that sexual differentiation does not require GnRH signaling, suggesting that the neonatal endocrine events guiding sexual differentiation are GnRH-independent. Moreover, in addition to E₂'s well-described defeminizing effects on Kiss1 at birth, developmental E₂ also acts sometime before juvenile life to promote complete feminization of Kiss1 neurons, thereby enabling maximal Kiss1 expression in females later in adulthood.