The Glutamine-Glutamate Cycle Contributes to Behavioral Feminization in Female Rats

Successful sex differentiation leads to the expression of sex-specific mating behavior during adulthood, which is essential for species propagation in rodents. Sex differentiation includes brain masculinization and defeminization processes in male rodents and a brain feminization process in female rodents [1, 2]. Brain masculinization is a developmental cellular process that organizes brain regions permissive to the expression of male sexual behavior during adulthood. Brain defeminization causes the loss of the capacity to respond to the activational effects of ovarian hormones that induce the expression of female sex behavior. Both processes are distinct and necessary for the expression of male-specific mating behaviors, such as mounting, intromission, and ejaculation with a sexually receptive female rat [3‒5]. In contrast, brain feminization induces the capacity to respond to ovarian hormones in adulthood with lordosis, or female sexual receptivity.

Among the various theories that describe cellular mechanisms of masculinization and defeminization processes in rodents, the most widely accepted one posits that the brain is exposed to high levels of estradiol (E₂), which is aromatized from testicular androgen, during the perinatal critical period (between embryonic 18 and postnatal day [PN] 10) in rats [6, 7]. High E₂ exposure of the brain permanently increases dendritic spine number through increased prostaglandin-E2 release in neurons of the preoptic area (POA), contributing to brain masculinization [3]. It also changes dendritic spine number and astrocyte morphology through enhancing glutamate (Glu) and GABA release, respectively, in the medial basal hypothalamus (MBH), which are associated with brain defeminization in rats [4, 8]. In contrast, due to the quiescence of ovaries during the critical period, it remains unclear how the female brain is organized in the absence of or under conditions of low E₂ milieus.

Since dendritic spines are the primary loci of excitatory synaptic input [9], and astrocytic membrane interactions with neurons are actively involved in regulating synaptic activities and consequently controlling animal behaviors [10], the finding that higher E₂ levels induce an increase in spines and morphology changes of astrocytes suggests that interactions between glutamatergic synaptic signaling and astrocytes may play a role in the processes of sex differentiation [11‒13].

The glutamatergic neurons release Glu, which is responsible for transmitting most of the excitatory synaptic signaling in the brain. After the synaptic release of Glu, it is recycled using transporters, such as the excitatory amino acid carrier-1 (EAAC-1) located on presynaptic terminals [14‒16]. Nevertheless, up to 80% of the extracellular Glu in the synaptic cleft is removed by astrocytes using Glu transporter-1 (GLT-1) [17, 18]. Inside astrocytes, Glu is metabolized into Gln by an astrocyte-specific enzyme, Gln synthetase (GS) [15, 19]. Gln can then be expelled from the astrocytes by system N transporters (SN-1) and taken up into glutamatergic or GABAergic neurons by different system A transporter subtypes (SA-1 or SA-2) [20‒22]. Inside neurons, Gln is converted back to Glu by glutaminase. The synthesized Glu can then be transported into synaptic vesicles for release. In GABAergic neurons, it can be further decarboxylated into GABA, which is then packaged into vesicles for release [23, 24]. This amino acid metabolic pathway between astrocytes and neurons is named the Gln-Glu cycle (GGC; summarized by the glutamatergic synapse in Fig. 8).

Since the quantal content of synaptic vesicles is in equilibrium with the intra-terminal concentration of neurotransmitters [25, 26], the supply of Gln from astrocytes rapidly reflects the intra-terminal concentrations of Glu and consequently the Glu content in synaptic vesicles, and can regulate synaptic plasticity. This has been reported in the GGC-regulated GABAergic synapses in the hippocampal CA1 neurons of adult rats [24] and may also exist in the GGC-regulated glutamatergic synapse in neurons of the hypothalamus of perinatal rats [12]. We have reported that female pups constitutively rely on astrocytic Gln supply for sustaining ∼75% of glutamatergic synaptic efficacy in neurons of the ventral lateral part of the VMH (vlVMH; included in MBH), while male pups may depend on it for maintaining ∼50% of glutamatergic synaptic plasticity and only at higher neuron activities [12].

Furthermore, perinatal interruption of GGC by inhibition of neuronal Gln transport leads to sex-specific alterations in glutamatergic synaptic transmission (GST) of vlVMH neurons and induces changes in the organization of vlVMH neurons toward their sexual counterpart during adulthood [13]. Most of the alterations occur perinatally and seem to persist into adulthood [13], which is similar to the changes in the morphology of neurons or astrocytes in response to perinatal high E₂ exposure that mediates sex differentiation [4, 8].

Taken together, these findings indicate that fundamentally sexual dimorphic phenotypes of GGC-regulated glutamatergic synapses exist in the vlVMH neurons of the perinatal brains, which can be permanently modified by perinatal interruption of GGC and therefore may be associated with sex differentiation. Given that (vl)VMH neurons′ activities regulate the expression of female sexual behavior [27‒29] and are involved in the brain defeminization process of perinatal male rats [4, 30, 31], and that sexually dimorphic vlVMH neurons can control distinct male and female mating behaviors [4, 27, 32], we hypothesized that maintaining the intact GGC in the brain perinatally may be necessary for the sexual differentiation of both sexes, which then leads to expression of sex-specific mating behavior during adulthood.

To test this, perinatal pups of both sexes were administered various inhibitors, either to directly kill astrocytes or to target various points of the GGC, while supplemented with exogenous Gln when astrocytic Gln synthesis was inhibited. They were then raised until adulthood for reproductive behavior testing. After the behavior tests, protein levels of the key GGC-associated enzymes in the MBH were analyzed using Western blot. For comparison, an open-field test was also conducted for anxiety-like behavior and locomotor activity in adult animals pretreated perinatally with a neuronal Gln uptake blocker.

A total of 176 male rats and 162 female rats were used in the study. All animal experiments and procedures were approved by the AAALAC-certified animal facility of Chang Gung University (CGU) Animal Care and Use Committee. Adult Sprague-Dawley male and female rats were purchased from Bio LASCO (Taipei, Taiwan). The animals were mated and reared in animal center of CGU and were maintained under 14:10 h of light-dark cycle conditions (lights on at 6:00 a.m. and lights off at 8:00 p.m.) with free access to food and water. Pregnant females close to their delivery date were isolated and checked every morning, and the day on which pups were first found was determined as postnatal day 0 (PN0). Newborn pups were sexed based on anogenital distance. There were typically 10–15 pups per litter. Only litters with at least 4–6 sex pairs were used in the study. There were therefore 2–3 males and 2–3 females per litter per treatment group in each experiment, depending on litter size and sex composition. Animals of both sexes were evenly represented in each treatment group across 3 litters in alpha-(methylamino) isobutyric acid (MeAIB) and 6-Diazo-5-oxo-L-norleucine (DON) studies, and 5 litters in L-methionine sulfoximine (MSO) and sodium fluoroacetate (fluoroacetate) studies in each age group. All treated pups were housed with their treated and untreated (if any) littermates and their mother until weaning. Each pair of pups received daily intraperitoneal (ip) or bilateral intracerebroventricular (icv) injections of various GGC blockers to interrupt different steps of the GGC, which includes inhibition of astrocytic GS by MSO (50–100 mg/kg, ip), blockade of neuronal Gln transport by MeAIB (5 mm/1 μL, icv), inhibition of neuronal glutaminase activity by DON (6 mm/1 μL, icv), direct killing of astrocytes with fluoroacetate (4–5 mg/kg, ip) or vehicle (distilled water [DW; as a control group; control] as a solvent for MeAIB and DON; 0.9% saline [as a control group; control] as a solvent for the other drugs). In some experiments, L-Gln (0.1 m/kg, ip) was coadministered with MSO. The respective vehicle and drugs were administered on PN0 and PN1 or PN5 and PN6 for two consecutive days for all behavioral studies, except for the MeAIB study in Figure 1 and Tables 1, 2, and 5, and the DON study in Figure 6, which were administered only for 1 day on either PN0 or PN5 (or PN6) for behavior experimentation. We performed icv injections by penetrating the skin and skull with a beveled 23 gauge 2 μL microsyringe in cryoanesthetized pups. The icv injection sites were at perpendicular points caudally between the coronal and sagittal suture (next to the bregma) and a lower depth of 2 mm on PN0-PN1 pups, as previously reported [5]; they were at ± 1.1 mm lateral to bregma and a lower depth of 2.4 mm from the skull on PN5–6 pups [33]. Each icv infusion volume was 1 μL over a period of 60 s. To reduce disturbances to the dam while handling the pups, the vehicles or drugs were administered on PN0 at least 8–12 h after birth. However, handling the pups on PN5–6 produced fewer interruptions to the dam than handling them on PN0–1, and could thereby prevent their being cannibalized by the dam. Since there were no statistically significant differences in the behavioral results mediated by MeAIB pretreatment between PN0 or PN0–1 pups and PN5 or PN5–6 pups (Tables 1-3), the two age groups were combined for some of the results of the MeAIB experiment (Fig. 1, 7; Table 4).

After the pups were treated with drug(s) or vehicle, they were reared until their body weight reached at least 210 g in female rats and 270 g in male rats, which occurred at approximately PN60–67. All animals underwent gonadectomy under isoflurane anesthesia to eliminate the production of endogenous hormones. A testosterone-containing silicon capsule (active area: 30 mm; id: 1.57 mm; od: 3.18 mm; A-M Systems #808200; Sequim, WA, USA) was then implanted subcutaneously (sc) in the nape of the neck of pretreated male rats to standardize circulating hormone concentrations to those in normal adult male rats [3, 34]. The animals were then maintained on a reverse 12 h light/dark cycle (lights off at 9:00 a.m. and lights on at 9:00 p.m.) for the following behavioral testing. They were allowed to acclimate to the new light-dark cycle for 1 week, then each animal was placed in the Plexiglas test chamber (W × L × H = 50 × 38 × 25 cm) for 20 min per day for environmental acclimation (1 week before the behavioral test on PN67–73). Animals received injections of E₂ (10 μg in 0.1 mL corn oil, sc) 24 and 48 h, and progesterone (1 mg in 0.1 mL corn oil, sc) 4 h before testing lordosis behavior. The hormone concentrations and treatment regimen used were shown to induce female sexual receptivity reliably in previous studies [3, 5, 34].

Behavioral tests were first conducted on animals on PN74–80, during the middle dark phase, with monitoring and recording via an infrared camera. Each male rat was placed in the testing chamber 15 min before the introduction of a hormone-primed female or hormone-primed castrated male rat. Each drug-treated animal was tested with either an animal of the opposite sex that had received the vehicle pretreatment (Fig. 1-7) or with a naïve male rat (Tables 3, 4). In some tests, both the animals had been treated with vehicle; the results of such tests were pooled and used as a control.

Male sexual behavior assessment included sexual motivation and copulatory behavior expression. The former examined variables of latency (time) to first mounting and latency to first intromission. The latter examined parameters of total number of mounts (mounts in the absence of intromission), total number of mounts with pelvic thrusts (i.e., intromissions), and total number of ejaculations during a 30-min testing period. All latencies were calculated from the onset of the testing period; mounting was defined as the placement of both forepaws on the female’s flanks; intromission was defined as the observation of penis licking after rhythmic pelvic movements immediately following a mount (additional movie files show this in more detail; see online suppl. Files 1, 2; for all online suppl. material, see https://doi.org/10.1159/000541102); ejaculation was defined as the observation of a pause for 2 to 3 s, with body stiffness and, occasionally, backward arching of the body during intromission.

Female sexual behavior was assessed by determining the frequency and strength of the expression of lordosis behavior in response to each mount of a male rat. Lordosis behavior was defined as four paws being grounded with the hind region being elevated off the floor of the test chamber (online suppl. File 1). We measured two variables, unless otherwise noted: (1) lordosis quotient (LQ), defined as the number of lordosis events during the first 10 mounts, multiplied by 100; and (2) lordosis score (LS), rating the receptiveness (behavior intensity) of a female to mounting by the male for the first 10 mounts. LS was classified into four grades: 0 (no response) to 3 (complete dorsiflexion of the spine in the lordosis posture) (additional movie files show these in more detail, see online suppl. Files 1–3) [35].

Defeminization behavior was evaluated based on the expression of both feminizing and aggressive behaviors in hormone-primed castrated male rats (Tables 3, 4). The indices we used to examine feminization behavior were LQ, LS, the total amount of lordosis and immobility (see below) during the 30 min testing period. Aggressive behaviors were assessed by counting mount attempts and confrontation behaviors expressed by hormone-primed castrated male rats during the 30 min testing period. Mount attempts were defined as raising both forepaws toward the naïve male’s back. Confrontation behavior was defined as turning the body over in response to the naïve male’s mount and raising the forepaws toward the naïve male. We calculated the percentage of confrontations out of mounts by dividing the number of confrontation behaviors by the total number of mounts and multiplying the result by 100. Moreover, approximately half of the hormone-primed castrated males (four out of eight animals in the vehicle-pretreated group) also occasionally exhibited immobility or a slowing of movement immediately following the expression of a confrontation behavior. These behaviors were categorized as “immobility” (Tables 3, 4).

Most of the definitions and sexual behavior parameters used were as described in previous studies [3, 5, 34]. The tests were conducted once per week for three consecutive weeks (three trials) and the observers of the behavior movies were blind to the treatment groups in all behavioral experiments.

The animals were sacrificed on the morning of the day after the third behavioral trial (PN89∼95). Their brains were removed and tissues of the slices (600 μm thick) containing MBH (∼2.0–3.2 mm posterior to bregma) and POA (∼0.12–1.32 mm posterior to bregma) [36] were obtained with a stainless steel needle with an internal diameter of 1 mm, as previous described [12]. The Western blotting protocol for quantification of the level of interested proteins was as previously reported [12]. Briefly, MBH or POA tissues obtained from adult animals after behavioral testing were homogenized in ∼40 µL of RIPA lysis buffer (Millipore, Temecula, CA, USA). The homogenate was subjected to centrifugation and the supernatant was then used for a BCA protein assay (Thermo Fisher Scientific, Waltham, MA, USA) to determine protein concentration. Each protein sample (20 µg protein per lane) was subjected to SDS-PAGE in a 10% acrylamide gel and transferred onto a polyvinylidene fluoride membrane. The membrane was blocked by incubation with 20% skim milk and then incubated with an antibody against GS (1:30,000), glutaminase (1:2,000), GAPDH (1:70,000), or tubulin (1:60,000) overnight. The membrane was incubated for 1 h with the appropriate horseradish peroxidase-conjugated secondary antibodies (1:20,000∼1:50,000), developed with an enhanced chemiluminescence horseradish peroxidase substrate, and placed against autoradiography film (see online suppl. File 4). Band intensity on the resulting autoradiograph was measured with the Image J software. Data are expressed unitless, as the ratio of target protein to GAPDH or tubulin signals.

Perinatal rats of both sexes on PN5 were administered with vehicle or MeAIB (5 mm, bilateral, icv) and reared until the age of sexual maturation. They then underwent gonadectomy and were subjected to steroid hormone(s) supplements as per the procedure for the sex behavior test. All preparation and testing timings (in the middle of the dark phase) were the same as that for the sex behavior test described above.

The open field area was 50 cm wide and 35 cm long, which was divided into 25 5 × 5-cm rectangular areas. Each animal was first placed into the center rectangle of the arena and followed by video recording for 10 min of a test period. The expression and time spent on the following behaviors were used as an index for assessing anxiety behavior: travelling in the arena (locomotion), in the corner rectangular area, in an area other than the corner rectangular area, freezing time, and the total rearing times. After each individual test, the open field arena was cleaned with a 75% alcohol solution after the removal of urine and feces.

Data are expressed as mean ± standard error of the mean. Prism 8 software (GraphPad, Boston, MA, USA) was used for all statistical analysis. The Shapiro-Wilk test was used to assess data for distribution normality, which determined whether parametric or nonparametric tests were used. If any data were not normally distributed, nonparametric analysis was performed. For the parametric analyses, p values were obtained using two-tailed Student’s t tests (MeAIB, fluoroacetate, and DON studies; Fig. 1, 3, 4, 6, and 7; Tables 1-4), one-way ANOVAs (part of the MSO studies; Fig. 2, 5), two-way ANOVAs (Tables 1-3), repeated measures three-way ANOVAs (online suppl. File 5), followed by Bonferroni’s test for multiple comparisons. For the nonparametric tests, p values were obtained from Mann-Whitney two-tailed tests (part of the MeAIB, fluoroacetate, and DON studies; Fig. 1, 3, 4, and 6; Table 4) or Kruskal-Wallis ANOVAs followed by Dunn’s test for multiple comparisons (part of the MSO studies; Fig. 2, 5). Statistical significance was assessed at p ≤ 0.05, p ≤ 0.01, p < 0.001, or p < 0.0001, unless otherwise noted.

MeAIB (catalog # M2383), MSO (catalog # M5379), L-Gln (catalog # 49419), E2 (catalog # E8875), progesterone (catalog # P0130), and testosterone (catalog # 86500) were obtained from Sigma-Aldrich (St. Louis, MO, USA). Sodium fluoroacetate (catalog # 31220) was acquired from Fluka Research Chemicals (Charlotte, NC, USA). Anti-glutaminase antibody (Ab; catalog # ab93434, RRID: AB_10561964) and anti-tubulin Ab (catalog # ab7291, RRID: AB_2241126) were purchased from Abcam Biochemicals (Cambridge, UK). Anti-GS Ab (anti-GS; catalog # MAB302, RRID: AB_2110656), anti-glyceraldehyde-3-phosphate dehydrogenase (GAPDH, catalog # MAB374, RRID: AB_2107445) Ab, goat anti-rabbit IgG secondary Ab (catalog # AP132P, RRID: AB_90264), and goat anti-mouse IgG secondary Ab (catalog # AP124P, RRID: AB_90456) were obtained from Millipore.

Blockade of neuronal Gln uptake by central administration of MeAIB (5 mm, icv) for 1 day on PN0 significantly reduced LQ and LS in females to 65.3% and 62.7% of control levels, respectively (t₁₀ = 2.24, p = 0.049 for LQ; t₁₀ = 2.34, p = 0.041 for LS, n = 6 animals per group in each set of conditions; Table 1), but it had no effect on copulatory behavior expression or sexual motivation in males (Table 2). MeAIB administration for 1 day on PN5 or PN6 reproduced the MeAIB-mediated decrease in LQ and LS in females that was observed for PN0 (to 65.0% and 51.8%, respectively, of control levels; t₁₀ = 2.41, p = 0.037 for LQ; t₁₀ = 3.1, p = 0.013 for LS, n = 6 animals per group in each set of conditions; Table 1) but also had no effect on sex behavior expression in males (Table 2). Since there was no statistically significant difference in the MeAIB-mediated reduction in expression of female sexual behavior between the two age groups (interaction of pretreatment and age: p = 0.88 and p = 0.57 for LQ and LS, respectively; n = 6 animals per group in each set of conditions, two-way ANOVA; Table 1) and MeAIB had no effect on the sex behavior of males in either age group (interaction of pretreatment and age: p = 0.97, p = 0.93, p = 0.74, p = 0.46, and p = 0.38 for total number of mounts, intromissions, ejaculation, latency to mount, and latency to intromission, respectively; n = 5–9 animals per group in each set of conditions, two-way ANOVA; Table 2), we pooled the data from the two age groups. Administration of MeAIB (5 mm) for 1 day on PN0 or PN5 (or PN6) decreased LQ to 65.2% of control levels (86.11 ± 4.11 vs. 56.11 ± 7.78, control- vs. MeAIB-pretreated group, respectively; t₂₂ = 3.41, p = 0.0025, n = 12 animals in each group) (Fig. 1f); it also decreased LS to 56.9% of control levels (2.16 ± 0.16 vs. 1.23 ± 0.18, control- vs. MeAIB-pretreated group, respectively, t₂₂ = 4.0, p = 0.0006; n = 12 animals in each group) (Fig. 1g; online suppl. File 3). In contrast, perinatal MeAIB administration had no effect on the copulatory behavior expression and sexual motivation of male rats ([44.6 ± 6.0 vs. 44.7 ± 6.1, p = 0.99, for the number of mounts; 1.5 ± 0.28 vs. 1.46 ± 0.32, p = 0.93, for the number of ejaculations, control- vs. MeAIB-pretreated group, respectively, n = 11–16 animals per group in each set of conditions, two-tailed t test (Fig. 1a, c); 32.5 ± 4.1 vs. 32.6 ± 4.4, for the number of intromissions, p = 0.74; 59.0 ± 13.7 vs. 64.2 ± 22 for latency to first mounting, p = 0.97; and 85.7 ± 22.7 vs. 105.8 ± 44.7, p = 0.85, for latency to first intromission; control- vs. MeAIB-pretreated group, respectively, n = 11–16 animals per group in each set of conditions, two-tailed Mann-Whitney test]) (Fig. 1b, d–e; online suppl. File 2).

Inhibition of astrocytic Gln synthesis with lower doses of MSO (50–80 mg/kg, ip) for two consecutive days also reduced female sexual behavior expression (decrease LQ to 73.2% of control levels, 90.61 ± 2.8 vs. 66.36 ± 7.14, control- vs. MSO-pretreated group, respectively, one-way ANOVA: F_{3, 34} = 4.85, p = 0.006, post hoc, p = 0.006 (t₃₄ = 3.58); decrease LS to 70.2% of control levels, 2.48 ± 0.09 vs. 1.74 ± 0.21, control- vs. MSO-pretreated group, respectively, one-way ANOVA: F_{3, 34} = 6.1, p = 0.002, post hoc, p = 0.004 (t₃₄ = 3.79); n = 11 animals per group in each set of conditions) (Fig. 2f, g). The MSO-mediated inhibition of female sexual behavior was partly rescued by coadministration of Gln (0.1 m/kg) with MSO, indicating the essential role of astrocytic Gln in feminization (to 95.9% of control levels, post hoc, p = 0.068 for LQ; to 100% of control levels, post hoc, p = 0.02 (t₃₄ = 3.15) for LS, one-way ANOVA; n = 7–11 animals per group in each set of conditions; Fig. 2f, g). Again, the MSO pretreatment had no effect on male sexual behavior (52.0 ± 5.3 vs. 37.5 ± 5.1, p = 0.32 for the number of mounts; 1.59 ± 0.3 vs. 1.64 ± 0.30, p = 0.97 for the number of ejaculations; control- vs. MSO-pretreated group, respectively, n = 8–13 animals per group in each set of conditions, one-way ANOVA; Fig. 2a, c; 41.2 ± 4.5 vs. 29.9 ± 4.3, p = 0.33 for the number of intromissions; 42.7 ± 7.3 vs. 95.3 ± 33.5, p = 0.095 for latency to first mounting and 49.4 ± 6.56 vs. 102.5 ± 32.7, p = 0.17 for latency to first intromission; control- vs. MSO-pretreated group, respectively, n = 8–13 animals per group in each set of conditions, Kruskal-Wallis ANOVA) (Fig. 2b, d, e). Moreover, co-application of Gln with MSO had no effect on the expression of male sexual behavior (37.5 ± 5.1 vs. 39.7 ± 9.5, for the number of mounts; 1.64 ± 0.30 vs. 1.50 ± 0.52, for the number of ejaculations; MSO- vs. MSO plus Gln-pretreated group, respectively; n = 8–13 animals per group in each set of conditions, post hoc, p > 0.99 in each set of conditions, one-way ANOVA, Fig. 2a, c; 29.9 ± 4.3 vs. 26.7 ± 5.9, for the number of intromissions; 95.3 ± 33.5 vs. 152.2 ± 67.9, for latency to first mounting and 102.5 ± 32.7 vs. 152.5 ± 67.8, for latency to first intromission; MSO- vs. MSO plus Gln-pretreated group, respectively; n = 8–13 animals per group in each set of conditions, post hoc, p > 0.99 in each set of conditions, Kruskal-Wallis ANOVA, Fig. 2b, d, e). The administration of Gln (0.1 m/kg) alone had no effect on the mating behaviors of either male or female rats (52.0 ± 5.3 vs. 49.2 ± 7.0, for the number of mounts; 41.2 ± 4.5 vs. 32.6 ± 3.3, for the number of intromissions; 1.59 ± 0.3 vs. 1.75 ± 0.15, for the number of ejaculations; 42.7 ± 7.3 vs. 32.5 ± 6.6, for latency to first mount and 49.4 ± 6.56 vs. 41.5 ± 5.6, for latency to first intromission, control- vs. Gln-pretreated group, respectively, n = 8–9 animals per group in each set of conditions, post hoc, p > 0.99 in each set of conditions, one-way ANOVA or Kruskal-Wallis ANOVA; Fig. 2a–e for male rats; 90.61 ± 2.8 vs. 83.33 ± 4.57 for LQ; 2.48 ± 0.09 vs. 2.39 ± 0.13 for LS; control- vs. Gln-pretreated group, respectively, n = 9–11 animals per group in each set of conditions, post hoc, p > 0.99 in each set of conditions, one-way ANOVA; Fig. 2f, g for female rats). The results indicate that perinatal MeAIB and MSO pretreatment consistently decreased both the frequency and strength of lordosis behavior in female rats, but had no effect on sexual motivation or the performance of copulatory behavior in male rats. Therefore, the results suggest that both astrocytic Gln and its transport into neurons are essential for the behavioral feminization of female rats but may play no role in the behavioral masculinization of male rats.

The direct destruction of astrocytes by sodium fluoroacetate (4–5 mg/kg, ip) reduced sexual behavior in both male and female rats (in female rats: decrease LQ to 71.3% of control levels, 87.0 ± 4.0 vs. 62.0 ± 5.4, t₁₈ = 3.73, p = 0.002; decrease LS to 72.3% of control levels, 2.38 ± 0.17 vs. 1.72 ± 0.16, t₁₈ = 2.82, p = 0.012, control- versus fluoroacetate-pretreated group, respectively; n = 10 animals per group in each set of conditions (Fig. 3f, g); in male rats: to 47.9% of control levels, 49.6 ± 5.5 vs. 23.8 ± 5.1, t₁₆ = 3.44, p = 0.003 for total number of mounts; to 47.5% of control levels, 39.6 ± 4.7 vs. 18.8 ± 4.1, t₁₆ = 3.36, p = 0.004 for total number of intromissions, control- vs. sodium fluoroacetate-pretreated group, respectively; n = 8–10 animals, two-tailed t test (Fig. 3a–b); to 37.9% of control levels, 1.58 ± 0.34 vs. 0.60 ± 0.27, p = 0.033 for the total number of ejaculations; to 575.4% of control levels, 45.4 ± 14.1 vs. 261.5 ± 123.9, p = 0.10 for latency to first mount; and to 605.2% of control levels, 55.0 ± 18.0 vs. 333.1 ± 174.8, p = 0.083 for latency to first intromission, control- vs. sodium fluoroacetate-pretreated group, respectively; n = 8–10 animals, two-tailed Mann-Whitney test (Fig. 3c–e). The data suggest that astrocytes are required for both behavioral feminization and masculinization of female and male rats, respectively.

To validate the effects of the GGC on the expression of mating behavior, perinatal rats of both sexes were administered with MeAIB (5 mm, icv) for two consecutive days and underwent sex behavior testing during adulthood.

The results for the pups that received MeAIB administration for 2 days were similar to those for the pups that received 1-day injections. Specifically, the frequency and strength of the sexual behavior of the females were reduced, with LQ at 66.7% of control levels (control- vs. MeAIB-pretreated group: 90.0 ± 4.6 vs. 60.0 ± 7.3; t₈ = 3.48, p = 0.008), LS at 64.4% of control levels (control- vs. MeAIB-pretreated group: 2.33 ± 0.17 vs. 1.5 ± 0.23; t₈ = 2.76, p = 0.025; n = 5 animals per group in each set of conditions; Fig. 4f, g). MeAIB did not affect the expression of mating behavior in males, with the total number of mounts, intromission, and ejaculations at 55.2 ± 4.7 versus 52.9 ± 2.2, 39.0 ± 4.0 versus 36.7 ± 4.4, and 1.25 ± 0.35 versus 1.73 ± 0.44, respectively, for the control versus MeAIB-pretreated groups (p = 0.4–0.72, n = 5–8 animals per group in each set of conditions; two-tailed t test; Fig. 4a–c). Similarly, there were no differences in latency to first mounting and latency to first intromission, which were 33.3 ± 6.5 versus 25.1 ± 9.7 and 40.0 ± 5.9 versus 25.5 ± 9.6, respectively, for the control- versus MeAIB-pretreated groups (p = 0.18–0.22, n = 5–8 animals per group in each set of conditions; two-tailed Mann-Whitney test; Fig. 4d, e). These results support the aforementioned findings that Gln transport into neurons is essential for brain feminization in female rats but plays little or no role in brain masculinization in male rats.

Higher doses of MSO (90–100 mg/kg) reduced the expression of sexual behavior in female rats, and the reduction in LQ and LS was partly rescued by coadministration of MSO with Gln. Injection of MSO reduced LQ to 73.5% of control levels (95.61 ± 1.72 vs. 70.3 ± 8.7, control- vs. MSO-pretreated group, respectively; p = 0.024, post hoc, p = 0.045, n = 11 animals per group, Kruskal-Wallis ANOVA; Fig. 5f). LQ was restored to 98.8% of control when Gln (0.1 m/kg) was coadministered with MSO (94.52 ± 2.32; post hoc, MSO vs. MSO plus Gln, p = 0.32, n = 7–11 animals per group, Kruskal-Wallis ANOVA; Fig. 5f). Injection of MSO also decreased LS to 69.3% of control levels (2.64 ± 0.05 vs. 1.83 ± 0.23, control- vs. MSO-pretreated group, respectively; Fig. 5g). LS was recovered to a score of 2.51 ± 0.12 when Gln (0.1 m/kg) was coadministered with MSO (F_{3, 34} = 5.39, p = 0.004; post hoc, control vs. MSO, p = 0.004 [t₃₄ = 3.78]; MSO vs. MSO with Gln, p = 0.0506 [t₃₄ = 2.8], n = 7–11 animals per group in each set of conditions, one-way ANOVA; Fig. 5g). These results support the aforementioned findings and suggest an essential role of astrocytic Gln in brain feminization in female rats.

In contrast, perinatal pretreatment with higher MSO concentrations had no effect on the expression of male copulatory behavior or motivation for the expression of copulatory behavior. For the control- versus MSO-pretreated groups, the number of mounts and intromissions were 55.3 ± 4.9 versus 41.0 ± 4.8 (p = 0.28) and 42.8 ± 4.8 versus 33.5 ± 6.6 (p = 0.47), respectively (n = 7–10 animals, one-way ANOVA; Fig. 5a, b), and the number of ejaculations, latency to first mount, and latency to first intromission were 1.67 ± 0.23 versus 1.03 ± 0.29 (p = 0.12), 32.6 ± 5.4 versus 151.4 ± 46.8 (p = 0.14), and 44.6 ± 7.5 versus 118 ± 23.6 (p = 0.18), respectively (n = 7–10 animals per group, Kruskal-Wallis ANOVA; Fig. 5c–e). Co-administration of Gln (0.1 m/kg) with MSO had no effect on any of the behavior parameters tested in males. For the MSO versus MSO with Gln-pretreatment groups, the number of mounts and intromissions were 41.0 ± 7.0 versus 41.0 ± 4.8 and 33.5 ± 6.6 versus 37.6 ± 4.2, respectively (n = 7–10 animals per group in each set of conditions, post hoc, p = 0.95–0.99, one-way ANOVA; Fig. 5a, b), and the number of ejaculations, latency to first mount, and latency to first intromission were 1.03 ± 0.29 versus 2.0 ± 0.43, 151.4 ± 46.8 versus 63.8 ± 35, and 118.0 ± 23.6 versus 70.3 ± 44.9, respectively (n = 7–10 animals per group in each set of conditions, post hoc, p = 0.16, p > 0.99 and p = 0.26 for the number of ejaculations, latency to first mount, and latency to first intromission, respectively, Kruskal-Wallis ANOVA; Fig. 5c–e).

The administration of Gln (0.1 m/kg) alone had no effect on the expression of mating behaviors in either males or females (Fig. 2a–e). Again, these results support the aforementioned findings that astrocytic Gln and its transport into neurons are essential for brain feminization in female rats. These results also suggest that astrocytic Gln may play a minor or no role in brain masculinization in male rats.

MeAIB administration in male pups reduces AMPA receptor- and NMDA receptor-mediated currents and alters the morphology of vlVMH neurons toward the female phenotype during adulthood [13]. This suggests that brain defeminization processes could be interrupted by GGC inhibition in male pups, leading to a decrease in the expression of sex behavior during adulthood. To test this hypothesis, we administered MeAIB to male pups at either PN0–1 or PN5–6 for two consecutive days and examined their expression of feminized and aggressive behavior during adulthood. MeAIB administration at PN0–1 or PN5–6 did not enhance the induced expression of lordosis or aggressive behavior in hormone-primed castrated male rats during adulthood (Table 3). Because there were no statistically significant differences between the results obtained from the two age groups (Table 3), we pooled the data for the two age groups. The combined results showed that perinatal MeAIB administration had no effect on any of the defeminization parameters tested (Table 4). These results suggest that the inhibition of neuronal Gln transport may not play a role in brain defeminization mediated by high levels of endogenous E₂ in perinatal male rats.

We tested whether neuronal glutaminase, which synthesizes Glu from Gln, acting downstream of the GGC, could modulate the expression of sex behavior of both sexes. To this end, perinatal rats of both sexes were centrally administered a neuronal glutaminase inhibitor, DON (6 mm, icv), to directly inhibit Glu synthesis on PN5. Pretreatment with DON reduced LQ and LS to 70.9% and 66.2% of control levels, respectively (LQ: 91.43 ± 4.82 vs. 64.83 ± 5.96, control- vs. DON-pretreated, p = 0.01, n = 7–8 animals per group, two-tailed Mann-Whitney test; LS: 2.63 ± 0.13 vs. 1.74 ± 0.12, control- vs. DON-pretreated, t₁₃ = 4.87, p = 0.0003, n = 7–8 animals per group, two-tailed t test; Fig. 6f, g). In contrast, perinatal DON pretreatment had no effect on the performance of copulatory behavior or motivation for copulatory behavior in male rats (44.2 ± 8.6 vs. 37.2 ± 5.5, p = 0.53 for the number of mounts; 36.4 ± 6.6 vs. 32.1 ± 4.1, p = 0.6 for the number of intromissions, 1.48 ± 0.36 vs. 1.22 ± 0.37, p = 0.66 for the number of ejaculations, control- vs. DON-pretreated group, respectively, n = 6–7 animals per group in each set of conditions, two-tailed t test; 41.5 ± 15.4 vs. 238.4 ± 175.1, p = 0.10 for latency to first mounting; 51.2 ± 16.6 vs. 247.1 ± 173.6, p = 0.10 for latency to intromission, control- vs. DON-pretreated group, n = 6–7 animals per group in each set of conditions, two-tailed Mann-Whitney test; Fig. 6a–e). These findings are consistent with our findings for MeAIB administration and support the hypothesis that GGC-regulated glutamatergic synaptic efficacy mediates behavior feminization in female rats, probably by modulating vesicular Glu content and basal Glu release in the perinatal pups [13].

To further investigate whether perinatal administration of GGC blocker had a long-term effect on the GGC-associated proteins, which in turn might lead to altering the expression of sexual behavior, the protein levels of two key enzymes of the GGC, glutaminase and GS, in the MBH and POA of the animals with perinatal MeAIB pretreatment and following behavioral tests were measured. Basal protein levels of glutaminase or GS in the MBH did not differ between adult male and female rats pretreated with vehicle (glutaminase/tubulin ratios: 0.72 ± 0.07 vs. 0.64 ± 0.06, males vs. females, p = 0.38; GS/tubulin ratios: 0.98 ± 0.05 vs. 0.94 ± 0.05, males vs. females; p = 0.38; two-tailed t test; n = 37–39 animals of each sex per group; figures not shown). Perinatal MeAIB injection increased protein levels of glutaminase in the MBH of adult female rats, whereas the pretreatment had no effect on that of male rats (glutaminase/tubulin ratio: 154 ± 14% of control levels, 0.44 ± 0.02 vs. 0.68 ± 0.07, control vs. MeAIB, t₂₄ = 3.51, p = 0.002, in females, two-tailed t test; 91 ± 10% of control levels, 0.73 ± 0.15 vs. 0.55 ± 0.06, control vs. MeAIB, respectively, p = 0.68, in males, two-tailed Mann-Whitney test; n = 12–14 animals of each sex per group; Fig. 7a, b). There was also a female-specific increase in glutaminase protein levels in response to MeAIB pretreatment compared to that of male animals after behavior testing (male vs. female percentage changes compared to control levels for respective sex: t₂₄ = 3.65, p = 0.0013, n = 12–14 animals, two-tailed t test; Fig. 7d). In contrast, perinatal MeAIB injections had no effect on GS levels in the MBH in either sex (GS/tubulin ratio: 0.92 ± 0.04 vs. 0.91 ± 0.06, control vs. MeAIB, in males; 0.79 ± 0.07 vs. 0.89 ± 0.11, control vs. MeAIB, in females; n = 14 animals of each sex per group; Fig. 7a, c, d). Taken together, these results support our previous findings [12, 13] that interruption of the GGC via perinatal MeAIB administration specifically alters GST and synaptic organization in the hypothalamus of female rats. This interrupts the feminization processes of female rats but has little to no effect on the behavior masculinization or defeminization of male rats during brain development.

On the other hand, basal protein levels of glutaminase in the POA did not differ between adult male and female rats pretreated with vehicle (glutaminase/tubulin ratios: 0.79 ± 0.04 vs. 0.86 ± 0.06, males vs. females, p = 0.35, n = 33–35 animals, two-tailed t test; figure not shown), despite significant differences in basal GS levels in the POA between adult male and female rats (GS/tubulin ratios: 1.01 ± 0.05 vs. 0.85 ± 0.04, males vs. females, t₆₆ = 2.7, p = 0.009, n = 33–35 animals; figure not shown). There was no change in glutaminase or GS levels in the POA in either MeAIB-pretreated males or females; the glutaminase/tubulin ratios were 0.64 ± 0.06 versus 0.58 ± 0.05 (p = 0.41) and 0.62 ± 0.06 versus 0.67 ± 0.04 (p = 0.55) in males and females, respectively (n = 11–12 animals of each sex per group, two-tailed t test; Fig. 7e, f, h), and the GS/tubulin ratios were 0.88 ± 0.06 versus 0.99 ± 0.07 (p = 0.25) and 0.93 ± 0.05 versus 0.92 ± 0.05 (p = 0.93) in males and females, respectively (n = 12 animals of each sex per group, two-tailed t test; Fig. 7e, g, h). The results in POA agree with the finding that MeAIB pretreatment had no effect on masculinization behavior expression in adults.

Since disruption of GGC is more involved in the development of psychiatric disturbances in female than in male rats [37], we tested whether perinatal administration of MeAIB could affect the locomotor activity and anxiety-like behaviors of both sexes during adulthood. There were no sex differences in the locomotor activity and expression of anxiety-like behaviors (p = 0.08 for total time spent on locomotion; p = 0.78 for total rearing times; p = 0.62 for time spent on the corner; p = 0.09 for the time spent on any rectangular area other than corner; n = 7 animals in each sex group, two-tailed t test) (Table 5). Perinatal MeAIB administration had no effects on locomotor activity and expression of anxiety-like behaviors compared to their respective control group during adulthood (p = 0.34 and p = 0.15 for the total time spent on locomotion; p = 0.68 and p = 0.18 for the total rearing times; p = 0.85 and p = 0.84 for the time spent on the corner; p = 0.42 and p = 0.21 for the time spent on any rectangular area other than corner; male and female groups, respectively; n = 7 animals in each group in each sex, two-tailed t test) (Table 5). Moreover, neither perinatal MeAIB pretreatment induced expression of any freezing behavior between the sexes or compared to their respective control group (Table 5).

Despite the well-established role of high E2 exposure in the POA and MBH (including VMH) mediating distinct cellular mechanisms for masculinization and being associated with the defeminization processes of the male brain [3, 4], the mechanisms underlying the feminization process in the absence of or with lower E2 levels in the default female rodent brain have not been fully elucidated. Using electrophysiological methods and identification of the morphology of neurons recorded from the specific small area of the hypothalamus that is associated with expression of female sex behavior, we have previously shown that sexually dimorphic phenotypes of GGC-regulated glutamatergic synaptic efficacy exist in the vlVMH neurons of perinatal rats, i.e., activity-dependent versus constitutive roles in the male and female pups, respectively [12]. Perinatal blockade of the GGC, achieved through the inhibition of neuronal Gln transport by MeAIB, induces sex-differential alterations in glutamatergic pre- and postsynaptic signaling of vlVMH neurons in adult rats. Additionally, it changes the synaptic organization of vlVMH neurons in adult rats, aligning them to the opposite sex [13].

Building upon these findings, the present study further demonstrates that disruption of the GGC at any point consistently decreased the expression of lordosis behavior in female rats. Conversely, the inhibition had no effect on brain masculinization or defeminization in males. Moreover, the effect of decreased mating behavior mediated by blockade of astrocytic Gln synthesis, MSO, in female rats was partly rescued perinatally when Gln was co-applied with MSO, indicating a key functional role of astrocytic Gln in mediating female sex differentiation for reproductive behavior expression. The long-term changes in perinatal GGC disruption were validated by the findings of a sex-specific increase in glutaminase protein in the MBH, but not POA, of MeAIB-pretreated animals following behavior testing. In contrast, perinatal administration of MeAIB had no effect on locomotor activity and anxiety-like behavior of animals.

The results of this study are consistent with our previous findings [12, 13] and further strengthen the hypothesis that the Gln provided by the GGC regulates GST of vlVMH neurons, which is essential for brain feminization in female pups, while playing little or no role in brain masculinization or defeminization in male pups. This is the first study to comprehensively examine the functional roles of GGC on sex differentiation of reproductive behaviors in perinatal rats.

Fluoroacetate is a toxin that specifically blocks the tricarboxylic acid cycle, thereby blocking the supply of ATP to glial cells and directly killing them [38, 39]. We also found that animals of either sex with perinatal fluoroacetate pretreatments reduced body weights compared with their respective saline-treated group during development, although there were sex-specific reductions in body weight during brain development (see subheading “Perinatal astrocytes are necessary for both behavior masculinization and behavior feminization in male and female pups, respectively” below and online suppl. File 5). MeAIB is a specific competitive and reversible inhibitor of neuronal system A (SA) transporters (Fig. 8) that are normally saturated with extracellular Gln (∼0.4 mm) [40, 41]. Application of MeAIB roughly doubles basal extracellular Gln concentrations in the brain of awake rats, indicating blockade of most neuronal Gln transport [42]. Perinatal acute application of MeAIB decreases glutamatergic synaptic efficacy of vlVMH neurons under higher synaptic activities in brain slices of both sexes and produces long-term effects on the glutamatergic pre- and postsynaptic machinery of male and female adult rats [12, 13]. This study further demonstrates that these may interrupt the processes of brain sex differentiation, particularly for female rats.

GS is an enzyme specifically expressed in astrocytes. MSO can be phosphorylated on the GS to form MSO phosphate, which binds tightly to the active site of GS and permanently inhibits GS activity [43]. Intraperitoneal injection of MSO (100–150 mg/kg) induces an inhibition of 30–70% of GS activity in PN10 pups; it also decreases protein expression of GABA_A receptor subunits in the adult brain [44, 45], which indicates an interruption in the balance established by inhibitory and excitatory synapses.

DON inhibits glutaminase activity by binding competitively against Gln for the active site of glutaminase [46, 47]. Brain slices preincubated with DON (6 mm) in the presence Gln have decreased Glu and GABA content in the synaptic terminals, indicating that DON specifically inhibits glutaminase, and that the glutaminase reaction is important for Gln to act as a precursor for the synthesis of amino acid neurotransmitters [48].

Perinatal administration of fluoroacetate reduced body weights and the expression of mating behaviors in male and female rats, albeit with sex-specific effects on body weight during development (interaction of age, sex, and pretreatment, F_{6, 276} = 4.92, p < 0.0001, repeated measures three-way ANOVA; online suppl. File 5A). Compared to the saline-pretreated group, the body weight of fluoroacetate-pretreated males was reduced throughout all developmental stages until PN40, while that of females was reduced only before weaning on PN10 and PN20 (online suppl. File 5B, C). These results highlight sex-specific differences in astrocyte-associated energy expenditure during development in rats, which may be associated with the GGC and sex differentiation. Since astrocytes can produce and release prostaglandin-E2 (PGE2) and Gln, the former serve as an E₂-mediated downstream molecule to masculinize the male brain [3], whereas the latter may serve as a constitutive precursor for neuronal Glu synthesis in neurons of vlVMH of female rats during the critical period [12, 13]. Thus, it is reasonable to conclude that perinatal fluoroacetate administration decreased most measured indices of mating behavior in both sexes, although the possibility that fluoroacetate-induced decreased male behavior expression by killing microglia cannot be excluded [49].

Synaptic vesicular Glu content appears to be dynamically equilibrated with the availability of Gln, and subsequently Glu, in the presynaptic terminal cytoplasm of vlVMH neurons of female pups, as previously reported in the GABAergic synapses [24]; as a result, perinatal blockade of Gln availability by either MeAIB (5 mm) or MSO (1.5 mm) rapidly reflects decreases in Glu content in synaptic vesicles and release recorded from brain slices [12]. The altered synapses may persist partly into adulthood during female brain development [13]. As the astrocytic Gln normally contributes to ∼50% and ∼75% of GSTs in female pups under basal and higher neuron activity, respectively [12], perinatal MeAIB (5 mm) pretreatment leads to a compensatory increase in postsynaptic AMPA receptor number, probably due to an increase in spine numbers and the process length of vlVMH neurons during development [13].

The present finding that MeAIB (5 mm) pretreatment increased neuronal glutaminase protein in the MBH of adult female rats reflects the previous findings of increased process lengths of vlVMH neurons [13]. The present finding, of perinatal blockade of the GGC at any point consistently decreasing expression of lordosis frequency and strength, agrees with our previous findings and further support the hypothesis that constitutively GGC-dependent glutamatergic synapses play a pivotal role in the functional development of hypothalamic neurons, particularly for female brain feminization. Perinatal interruption of the GGC therefore disrupts brain feminization processes and leads to brain defeminization of female rats.

Moreover, the decrease in female sex behavior under GGC disruption appears to not be due to an increase in anxiety-like behavior or a decrease in locomotor activity, as no such effects were found in the open-field test in response to perinatal MeAIB administration in either sex. Nevertheless, given the variation in litter sizes and numbers of sex pairs, litter and cohort effects cannot be excluded as confounding variables. On the other hand, since E₂ can upregulate two key GGC enzymes, GS and glutaminase [50, 51], further studies are required to determine whether developmental changes of the GGC-dependent glutamatergic synaptic phenotype exist and its roles for behavior feminization, as aromatized E₂ is necessary for feminization during prepuberty [52].

Neither glutaminase nor GS protein levels were altered in the POA of MeAIB-pretreated animals after behavior testing (Fig. 7e–g, h). The lack of an effect of MeAIB or any dose of MSO on the performance of sex behavior by the males suggests that GGC-associated synaptic plasticity may play little or no role in mediating POA-associated brain masculinization [3]. Another potential confounder that should be considered is that MSO, especially at high doses, perturbs not only glutamatergic but also GABAergic signaling, and may lead to ammonia toxicity [44, 45]. In our study, we noted occasional instances of animal death upon administration of a single high dose of MSO (90–100 mg/kg).

On the other hand, while perinatal administration of MeAIB mediates decreases in both postsynaptic AMPA receptor- and NMDA receptor-mediated currents in male rats, which could interrupt the process of brain defeminization [4, 13], an increase in the probability of basal synaptic Glu release, which contributes to brain defeminization processes, also occurs in vlVMH neurons recorded in the presence (or pretreatment) of MeAIB in brain slices obtained from male rats [4, 12, 13]. Thus, it is not surprising to find that the brain defeminization process seems to have been preserved in the MeAIB-pretreated hormone-primed castrated male rats.

Nevertheless, it is interesting to observe that approximately half of the hormone-primed castrated male rats occasionally exhibited reduced mobility or immediate immobilization after a confrontation behavior in response to a mount (four out of eight animals tested; 2/3 and 2/5 of animals from PN0–1 and PN5–6 control groups, respectively; Table 3), suggesting the expression of a female proceptive behavior. None of the animals (0/3) pretreated with MeAIB at PN0–1 expressed the immobility behavior, whereas the percentage of animals pretreated with MeAIB at PN5–6 expressing immobility behavior (2/5) was the same as that of the control group (2/5). Moreover, male pups pretreated with MeAIB at PN0–1 expressed more aggressive behavior as they tended to mount naïve male rats (p = 0.052 compared to control, n = 3; Table 3). Although the amount of time spent in immobility and the number of mount attempts did not differ significantly between the pretreatment groups, nor differed significantly between different age groups (Table 3), these observations are consistent with our hypothesis that the GGC may be essential for the default progression of brain feminization. As a result, blockade of the GGC by MeAIB may enhance processes of brain defeminization in male pups at PN0–1. Further research is needed to clarify the role of the GGC on brain defeminization processes in male pups since the testosterone surge that occurs in male rats on PN0, in the first few hours after birth, has a significant impact on sexual differentiation [53, 54].

Moreover, male pups show higher protein levels of glutaminase and GS compared to female pups in the MBH [12], which suggests that male pups may have a higher demand for the GGC-regulated synaptic plasticity than female pups. The administration of MeAIB, in addition to inhibiting SA-2 on the glutamatergic neurons, also blocks SA-1 located on the GABAergic neurons [22], which may act as another source of excitatory inputs during the perinatal period. Therefore, an alternative source of excitatory drive from “non-GGC-regulated” depolarizing GABA may be upregulated to compensate for the loss of excitatory drive, thus maintaining the excitatory synaptic strength required for masculinization and defeminization. Further studies on the roles of GGC in regulating GABAergic synaptic transmission in vlVMH neurons of both sexes of perinatal rats are needed to elucidate this question.

Although it remains unclear whether the decreased behavioral effects of GGC blockers were due to direct action on vlVMH neurons or indirect action of other cells that project to vlVMH neurons, the results of inhibition at each point of the GGC, particularly the inhibition of neuronal Gln transport by MeAIB, a key step of the GGC, consistently inhibited feminized behaviors in females. These results align with our previous findings showing that perinatal disruption of the GGC causes acute and long-lasting decreases in synaptic Glu release in female rats. The findings uncover the functional role of GGC-regulated glutamatergic synaptic vesicular content and release, which may contribute to sex differentiation of behavior expression in female rats. The findings also suggest that behavioral defeminization could occur in the absence of high levels of E₂ exposure to the brain in female rats during the perinatal period. This study may help fill in the gap in knowledge of sex differentiation and shed light on the synaptic pathophysiological mechanisms of sex-based developmental diseases.

The intact GGC is essential for behavioral feminization in female rat pups but may play a minor or no role in behavioral masculinization and defeminization in male pups during brain development.

The authors would like to thank Miss Shin-Ting Chen for watching the movies, Professor Min-Chi Chen’s suggestions on the statistical analysis, Professors Hung-Li Wang, Jin-Chung Chen, Sheng-Chieh Hsu, and the Neuroscience Research Center of Chang Gung Memorial Hospital at Linkou, Taiwan, for their support.

The Institutional Animal Care and Use Committee of Chang Gung University (No. CGU108-110) approved all procedures following the guidelines in the Guide for Laboratory Animal Facilities and Care as endorsed by the Council of Agriculture, Executive Yuan, ROC.

The authors declare no conflicts of interest.

This study was supported by grant MOST 106-2320-B-182-010 and 110-2320-B-182-014 (to S.L.L.) from the Ministry of Science and Technology, Taiwan, ROC; and by grants CMRPD1K0191, CMRPD1K0571, BMRPB35 (to S.L.L.) and CMRPD1M0821 (to S.L.L. and R.S.C.) from the Chang Gung Medical Foundation, Taiwan, ROC.

S.L.L. acquired funding for the experiments, designed and performed the experiments, analyzed the data and prepared, edited figures, and the manuscript. R.S.C. acquired funding for the experiments, edited the manuscript. S.L.L. and R.S.C. approved the final manuscript.

All data generated or analyzed during this study are included in this article and its supplementary material files. Further inquiries can be directed to the corresponding authors.