Male Fmr1 KO and control animals on a C57BL/6J background were generated from heterozygous and control breeder pairs in-house. Animals were weaned at 21 days of age. Genotyping was performed from tail biopsies as previously described (Qin et al., 2002). Animals were group housed in a climate controlled facility with access to food and water ad libitum. All procedures were performed in accordance with the National Institutes of Health Guidelines on the Care and Use of Animals and approved by the National Institute of Mental Health Animal Care and Use Committee.

At 21 days of age (P21), males in each litter were separated such that half received vehicle treatment and the other half received rapamycin treatment. Rapamycin was encapsulated with Eudragit (Rapamycin Holdings, San Antonio, TX, United States) and incorporated into mouse chow at 11 mg/kg food (Purina LabDiet, St. Louis, MO, United States). Based on how much food an average mouse consumes per day (4g; Tremblay et al., 2007), P21 animals received 4.4 mg/kg/day. Once the animals reached adult weight around 60 days of age, we estimate that the animals received 1.75 mg/kg/day. This dosage was based on a previous dosage used to reverse social behavior deficits in a mouse model of TSC (Reith et al., 2012).

Beginning at 60 days of age, animals were subjected to a battery of behavior tests with tests spaced 1 week apart. They were conducted from the least stressful to the most stressful in the following order: sleep testing, open field, zero maze, social behavior, and passive avoidance.

Sleep testing was conducted by means of a home-cage monitoring system (Columbus Instruments, Columbus, OH, United States) as previously described (Saré et al., 2017). Briefly, animals were individually housed and recorded for 72 consecutive hours. To eliminate the effect of habituation to the home-cage environment and single housing (Saré et al., 2017), we analyzed sleep duration from the last 48 h of recording. An animal was considered asleep if it had 40 s of consecutive inactivity. The percent time asleep was calculated separately for the light (inactive) and the dark (active) phases.

Activity and anxiety were assayed by means of photobeam detection in a novel open field environment (Coulbourn Instruments, Whitehall, PA, United States) as previously described (Saré et al., 2015). Briefly, total horizontal distance moved was recorded over a 30-min session in 5 min epochs as a measure of total activity. The ratio of horizontal distance traveled in the center (more than 1.91 cm away from the arena walls) to total distance was calculated for each epoch as an inverse measure of anxiety.

Anxiety was assayed by means of a zero maze (Med Associates, Fairfax, VT, United States). Animals were placed facing the open portion of the zero maze and allowed to explore the maze for 5 min. The times spent in the open and closed portions of the maze were determined. An animal was considered to be in a portion of the maze if both forepaws crossed into that portion. If an animal fell off the maze, it was eliminated from the analysis.

Social behavior was assayed by means of a 3-chambered apparatus as previously described (Saré et al., 2016). The assay had three consecutive parts. First, the animal was allowed a 5 min habituation period to the empty chamber. Animals demonstrating a side preference of more than 3 min were eliminated from the test. Second, in the sociability phase, an age-matched stranger mouse was placed in a sociability enclosure (Noldus Information Technology, Inc., Leesburg, VA, United States) in one chamber. An empty sociability enclosure was placed in the opposite chamber. These locations were alternated between mice to avoid a side bias. The test animal was placed in the center chamber and allowed to freely explore all chambers for 5 min. Times spent sniffing the sociability enclosures were recorded by subsequent video analysis. Sniffing was analyzed by means of the TopScan software (Clever Systems, Reston, VA, United States). Parameters were set to define sniffing as the animal’s nose directed toward the enclosure and within 20 mm of the enclosure. Third, in the social novelty phase, a novel stranger mouse was added to the previously empty sociability enclosure. The test animal was once again allowed to explore for 5 min, with measures recorded as before.

Passive avoidance was conducted as previously reported (Saré et al., 2016). Briefly, habituation was conducted on Day 1. The animals were placed in the lighted chamber of the passive avoidance shuttle box with the door to the dark chamber closed. After 30 s, the door to the dark chamber opened. Once the animal entered, the test was ended. On Day 2, the animal was subjected to two training sessions. The animal was once again placed in the lighted chamber with the door to the dark chamber closed. After 30 s, the door was opened. Once the animal entered the dark chamber, it received a footshock (0.3 mA, 1 s). The animal was left in the dark chamber for 15 s and then moved to a holding cage for 120 s and was then placed in the lighted chamber and the process was repeated. On Day 3 and 24 h after the training session, animals were placed in the lighted chamber with the door to the dark chamber closed. The door was opened and the latency to enter the dark chamber was recorded with a cutoff of 570 s.

In separate groups of animals, we conducted audiogenic seizure testing as previously reported (Miano et al., 2008). Testing began at P30. Animals were placed in a sound attenuating chamber with a viewing window (Med Associates, Fairfax, VT, United States). A siren (130 dB) sounded (Wal-Mart, Bentonville, AR, United States) for 5 min while seizure activity was observed. A seizure was defined as wild running, sometimes followed by myoclonic convulsions, sometimes followed by respiratory arrest. The frequency of each of these behaviors occurring in response to the tone was recorded.

Mice used for Western blotting had been tested for behavior with the exception of passive avoidance and audiogenic seizure susceptibility. A week after the last behavioral test, mice were decapitated and the frontal cortex was rapidly dissected for total protein extraction. The tissue was weighed and homogenized with 5% (weight/volume) solution of Tissue Protein Extraction Reagent (Thermo Fisher Scientific, Waltham, MA, United States) with 1% Halt Protease and Phosphatase Inhibitor Cocktail (Thermo Fisher Scientific) and 1% EDTA (Thermo Fisher Scientific). The homogenate was centrifuged at 15,000 × g for 15 min at 4°C and the supernatant was collected. Protein extracts (10 μg) were treated with equal volume 2× Laemmli buffer, incubated at 70°C for 10 min, and run on a 4–15% Mini-PROTEAN TGX Stain-Free gel (Bio-Rad Laboratories, Hercules, CA, United States). The gel was activated under ultraviolet light, proteins transferred to a nitrocellulose membrane, incubated with primary antibody overnight at 4°C, followed by secondary antibody [goat anti-rabbit horseradish peroxidase-linked at 1:10,000 (Bio-Rad Laboratories)] for 1 h at room temperature. We employed the Stain-Free Technology (Bio-Rad Laboratories) to normalize blots to total protein. The membrane was imaged under Stain-Free to determine total protein for loading control. The membrane was then incubated in Clarity substrate and imaged by means of a chemiluminescent signal on a ChemiDoc MP Imager (Bio-Rad Laboratories). Primary antibodies were used at a 1:1000 dilution and were as follows: FMRP (Abcam 27455), p-mTOR (Cell Signaling 5536), mTOR (Cell Signaling 2983), p-p70S6K (Cell Signaling 9234), p70S6K (Cell Signaling 2708), p-S6 235/236 (Cell Signaling 2211), p-S6 240/244 (Cell Signaling 2215), S6 (Cell Signaling 2217), p-ERK (Cell Signaling 4370), ERK (Cell Signaling 7124), p-AKT Ser473 (Cell Signaling 4060), and AKT (Cell Signaling 9272). Values presented are relative to the mean of vehicle-treated controls.

Data are reported as means ± SEM. Data from the zero maze, passive avoidance tests, and Westerns blots were analyzed by means of two-way ANOVA with genotype (control, Fmr1 KO) and condition (vehicle, rapamycin) as between subjects variables. Sleep and open field data were analyzed by means of a mixed-model repeated measures three-way ANOVA with genotype (control, Fmr1 KO) and condition (vehicle, rapamycin) as between subjects variables and phase (sleep) or epoch (open field) as within subjects variables. In cases in which the interaction between or among variables was statistically significant, we compared groups and/or conditions by means of post hoc t-tests. Social behavior was analyzed by means of paired student’s t-tests to compare stranger vs. object (sociability) or familiar vs. novel (social novelty). Audiogenic seizures were analyzed by means of a Fisher’s exact test comparing the effect of treatment in Fmr1 KO animals. Effects of p ≤ 0.05 are considered statistically significant and are denoted with a “^∗.” Effects of 0.10 ≥p > 0.05 are also reported here and are denoted with a “∼.” A table reporting the F-values and corresponding p-values for interactions and main effects are listed for the Western blots (Table 1) and all behavior tests analyzed with ANOVA (Table 2).

We analyzed lysates of frontal cortex from vehicle- and rapamycin-treated Fmr1 KO and control mice. FMRP expression was absent in all Fmr1 KO mice regardless of treatment (Figure 1A). In Figure 1, we present the results of the effects of genotype and rapamycin treatment on the phosphorylated forms (Figures 1B,E,H,K,N,Q), total (Figures 1C,F,I,L,O,R) and the ratio of phosphorylated to total (Figures 1D,G,J,M,P,S) for the signaling molecules measured. We did not find either an effect of genotype or of treatment on p-mTOR (Figure 1B and Table 1). We also did not find either an effect of genotype or treatment on p-p70S6k (Figure 1E and Table 1). We examined two phosphorylation sites on S6 (235/236 and 240/244), and in both cases, the genotype × treatment interactions were statistically significant (Table 1). Post hoc t-tests indicate that vehicle-treated Fmr1 KO mice had significantly higher p-S6 than vehicle-treated controls (p = 0.010 for 235/236) (p = 0.041 for 240/244). For both sites, rapamycin reduced p-S6 levels in Fmr1 KO mice (p = 0.009 for 235/236) (p = 0.024 for 240/244), but had no effect on controls (Figures 1A,H,K). To examine the mTORC2 pathway, we also found a statistically significant genotype × treatment interaction for p-AKT Ser473 (Table 1). Post hoc t-tests indicate that vehicle-treated Fmr1 KO animals had increased p-AKT relative to vehicle-treated controls (p = 0.02). Rapamycin reduced p-AKT levels in Fmr1 KO mice (p = 0.013), but had no effect on controls (Figures 1A,N). We also examined p-ERK. Although there was no genotype × treatment interaction, we did detect a statistically significant main effect of treatment for both p-ERK and total ERK. Regardless of genotype, rapamycin treatment decreased p-ERK and ERK levels (Figures 1A,Q,R and Table 1).

Given that p-S6 activity was higher in Fmr1 KO mice and reduced by rapamycin treatment, we hypothesized that rapamycin treatment would ameliorate the behavioral phenotypes in Fmr1 KO mice.

To examine hyperactivity, we conducted open field testing. We found a statistically significant main effect of genotype regardless of treatment, indicating that Fmr1 KO mice were hyperactive compared to controls, as previously reported (Liu et al., 2011; Figure 2 and Table 2). There was also a near statistically significant treatment × epoch interaction (p = 0.054) suggesting that rapamycin may enhance reactivity to the novel environment, regardless of genotype.

To examine anxiety, we analyzed the ratio of distance traveled in the center of the open field to total distance traveled. The main effect of epoch was statistically significant (p < 0.001) indicating that regardless of genotype or treatment, mice became less anxious as the test progressed (Figure 3A and Table 2). Main effects of genotype and treatment approached statistical significance suggesting that anxiety may be lower in Fmr1 KO mice regardless of treatment and that in both genotypes, rapamycin treatment tended to increase anxiety-like behavior.

As another measure of anxiety-like behavior, we tested animals in the zero maze (Figure 3B). In this test, the main effects of both genotype and treatment were statistically significant (Table 2). Zero maze results are in accord with open field behavior. Our results show that regardless of treatment, Fmr1 KO mice spent more time in the open arms indicating that Fmr1 KO mice are less anxious compared to controls as previously reported (Liu et al., 2011). Moreover, regardless of genotype, rapamycin decreased the time spent in the open arms indicating that rapamycin increased anxiety-like behavior in both genotypes.

To examine a form of fear learning, we tested mice on passive avoidance (Figure 4). We found a statistically significant main effect of genotype (p = 0.028) (Table 2), indicating that regardless of treatment, Fmr1 KO mice had a lower latency to enter the dark side. The main effect of rapamycin treatment was not statistically significant indicating that rapamycin did not enhance performance on this test.

Sleep in the active and inactive phases was assessed by home-cage monitoring. We found a nearly statistically significant genotype × phase interaction (p = 0.062) (Table 2). This is consistent with previous work indicating that decreases in sleep duration occurred primarily in the inactive phase in Fmr1 KO (Saré et al., 2017). We also found a statistically significant main effect of treatment (p = 0.047) (Table 2). These results indicate that rapamycin treatment reduced sleep duration across genotypes and phases (Figure 5).

To examine social behavior, we used the three-chambered apparatus to assess sociability and preference for social novelty. In the test of sociability, vehicle (p = 0.0003)- and rapamycin (p = 0.005)-treated control mice and vehicle-treated Fmr1 KO mice (p = 0.02) had a significant preference for the mouse compared to the object, but Fmr1 KO mice on rapamycin treatment did not (Figure 6A). This indicates that rapamycin impaired sociability in Fmr1 KO mice.

In the test of preference for social novelty, only the control mice on vehicle treatment had a significant preference for the novel mouse (p = 0.003) (Figure 6B). This was abolished with rapamycin treatment. Additionally and as previously reported (Liu et al., 2011), Fmr1 KO animals did not show a preference for the novel mouse. This was not changed with rapamycin treatment (Figure 6B).

We did not observe audiogenic seizures in either vehicle- or rapamycin-treated control mice (Figures 7A,B). Vehicle-treated Fmr1 KO mice had a 28% seizure incidence (11% wild running, 11% myoclonic convulsions, and 6% respiratory arrest) (Figure 7C). Rapamycin-treated Fmr1 KO had a 37% seizure incidence (11% wild running, 16% myoclonic convulsions, and 11% respiratory arrest) (Figure 7D). There were no statistically significant differences between vehicle and rapamycin-treated Fmr1 KO mice.