All procedures were reviewed and approved by the Association for Assessment and Accreditation of Laboratory Animal Care and the Institutional Animal Care and Use Committee of the Yonsei University Health System (permit number: 2018-0110, 2019-0336, and 2020-0054). All procedures were in accordance with the guidelines of the National Institutes of Health's Guide for the Care and Use of Laboratory Animals. These regulations, notifications, and guidelines originated and were modified from the Animal Protection Law (2008), Laboratory Animal Act (2008), and Eighth Edition of the Guide for the Care and Use of Laboratory Animals (NRC 2011). Mice were provided food and water ad libitum under a 12-h light/dark cycle, according to animal protection regulations. They were sacrificed at 8 weeks after the special housing conditions under anesthesia induced by intraperitoneal injection of ketamine (100 mg/kg) and xylazine (10 mg/kg) anesthesia by intraperitoneal injection. All efforts were made to minimize animal suffering.

At 6 weeks of age, a total of 104 (52 male and 52 female) ICR/CD-1 mice were randomly housed in either standard conditions (SC, n = 26 per sex) or an enriched environment (EE, n = 26 per sex). Additionally, a total of 70 mice underwent ovariectomy [OVX group, n = 24; OVX + Estradiol (E2) group, n = 15; OVX + EE group, n = 14; and OVX + E2/EE, n = 17]. The subcutaneous injection of E2 treatment (5 μg E2/corn oil) was administered every 3–4 days. The treatments (E2 and/or EE) for each group lasted until 14 weeks of age, as previously described (Seo et al., 2018). All mice were fed a normal chow diet, and vaginal cytology was conducted for all female mice immediately before sacrifice, as previously described (Byers et al., 2012).

Dual Energy X-Ray Absorptiometry (DXA) is a method to assess in vivo body composition in humans and animals, and can differentiate between fat and non-fat tissues. DXA was performed at 14 weeks of age to determine the whole-body composition of each group of mice. Each mouse was anesthetized with a mixture of xylazine and ketamine, transferred, and correctly positioned in the DXA chamber, and bone mineral content, fat, and lean body mass were measured.

All mice went through fasting process for 5–6 h before blood collection. Blood was drawn into BD Microtainer^® blood collection tubes (USA). Blood samples were centrifuged for 10 min at 300 g, and serum supernatants were collected and transferred to other tubes. Total protein, total cholesterol, high-density lipoprotein, triglyceride, and lactate dehydrogenase levels were measured using DRI-CHEM 4000i (Fuji).

To investigate the effect of EE on metabolism and physiology under the influence of estrogen, a total of 12 mice (n = 3 per group) for the normal model and a total of 22 OVX mice (OVX group, n = 6, OVX + E2 group, n = 6; OVX + EE group, n = 5, OVX + E2/EE, n = 5) were housed in metabolic automatic cages (Phenomaster, TSE systems) at 14 weeks of age. After 2 days of acclimation, measurements of energy expenditure, respiratory exchange ratio, and core temperature were taken at 9-min intervals for a total of 5 days.

βHB levels were measured in the blood serum and brain tissue extracts using the βHB assay kit from Sigma-Aldrich (MAK041-1KT) following the manufacturer's instructions. The values were obtained at 1:40 dilution and expressed as pmol/well, unless noted otherwise; each well contained a total volume of 200 μL.

Total RNA was prepared in the interested brain tissue lysates using TRIzol reagent (Invitrogen Life Technologies, Carlsbad, CA, USA) according to the manufacturer's instructions. A nanodrop spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) was used to confirm the quality and quantity of extracted RNA. Differentially expressed genes of interest from cerebral cortex and hippocampus were selected to validate by Quantitative Real-Time PCR (qRT-PCR). ReverTra Ace^® qPCR RT Master Mix with gDNA Remover (Toyobo, Osaka, Japan) was used to synthesize cDNA with total RNA. Then, 2 μl of cDNA in a total volume of 20 μl was used in the following reaction. The qRT-PCR was performed in triplicate on a Light Cycler 480 (Roche Applied Science, Mannheim, Germany) using the Light Cycler 480 SYBR Green master mix (Roche), with thermocycler conditions as follows: amplifications were performed starting with a 300 s template preincubation step at 95°C, followed by 45 cycles at 95°C for 10 s, 60°C for 10 s, and 72°C for 10 s. The melting curve analysis began at 95°C for 5 s, followed by 1 min at 60°C. The specificity of the produced amplification product was confirmed by the examination of a melting curve analysis and showed a distinct single sharp peak with the expected Tm for all samples. A distinct single peak indicates that a single DNA sequence was amplified during qRT-PCR. The detail sequence of the primers is listed in Supplementary Table 1. Primers were designed using the NCBI primer blast with the parameters set to a product of 150–200 bp within the region surrounding the identified translocation. The expression of each gene of interest was obtained using the 2^−ΔΔCT method. The expression level of each gene of interest was obtained using the 2^−ΔΔCT method. Target-gene expression was normalized relative to the expression of GAPDH and represented as fold change relative to the male control group.

Brain lysates were isolated from cerebral cortex and hippocampus. To assess the protein expression of MCT1, MCT2, MCT4, HDAC1, HDAC2, HDAC3, BDNF in cerebral cortex and hippocampus, total protein was extracted from all mice and dissolved in sample buffer (60 mM Tris–HCl, pH 6.8, 14.4 mM b-mercaptoethanol, 25% glycerol, 2% SDS, and 0.1% bromophenol blue; Invitrogen), incubated for 10 min at 70°C, and separated on a 10% SDS reducing polyacrylamide gel (Invitrogen). Twenty microgram of Protein samples were separated with SDS-polyacrylamide gel electrophoresis (PAGE) on a 4–12% gradient Bis-Tris gel and Tris-Acetate gel (Invitrogen, Carlsbad, CA, USA). The separated proteins were further transferred onto a 0.45 μm Invitrolon^TM polyvinylidene difluoride (PVDF) filter paper sandwich using a XCell II^TM Blot Module (Invitrogen, Life Technologies, Carlsbad, CA, USA). The membranes were blocked for 1 h in Tris-buffered saline (TBS) (10 mM Tris-HCl, pH 7.5, 150 mM NaCl) plus 0.05% Tween 20 (TBST) containing 5% non-fat dry milk (Bio-Rad, Hercules, CA, USA) at room temperature, washed three times with TBST, and incubated at 4°C overnight with the following primary antibodies; anti-MCT1 (1:500, Abcam), anti-MCT2 (1:200, Santa Cruz), anti-MCT4 (1:200, Santa Cruz), anti-HDAC1 (1:500, Santa Cruz), anti-HDAC2 (1:200, Santa Cruz), anti-HDAC3 (1:1,000, Santa Cruz), anti-BDNF (1:1,000, Abcam), and anti-β-Actin (1:5,000, Abcam). After washing the blots three times with TBST, the blots were incubated for 1 h with horseradish peroxidase-conjugated secondary antibodies (1:5,000; Santa Cruz, CA, USA) at room temperature. The proteins were further washed three times with TBST and visualized with an enhanced chemiluminescence (ECL) detection system (Amersham Pharmacia Biotech, Little Chalfont, UK). Using ImageQuant™ LAS 4000 software (GE Healthcare Life Science, Chicago, IL, USA), western blot results were saved into TIFF image files, and then the images and the density of the band were analyzed and expressed as the ratio relative to the control band density using Multi-Gauge (Fuji Photo Film, version 3.0, Tokyo, Japan). To normalize the values of all samples to account for band intensity, the average band intensity for each mouse group was first calculated. The samples were normalized to the group average of controls, and target protein expressions were normalized relative to the internal expression of β-Actin. The value of the male control group was set to 1 and was divided by the value of each individual mouse.

The brain tissues were frozen in Surgipath FSC 22 clear frozen section compound (Leica Microsystems) using dry ice and isopentane. The harvested brain tissues were cryosectioned into 16-μm-thick sections along the coronal plane, and immunohistochemical staining was performed as previously described. Eight weeks after EE, to confirm the endogenous expression of MAP2 (1:400, Abcam), GFAP (1:400, Neuromics), MCT2 (1:400, Santa Cruz), and MCT4 (1:400, Santa Cruz), the brain sections of the cerebral cortex and hippocampus were immunostained. The sections were incubated with Alexa Fluor^® 488 goat anti-rabbit (1:400, Invitrogen) and Alexa Fluor^® 594 goat anti-mouse (1:400, Invitrogen) secondary antibodies, and then covered with Vectashield^® mounting medium with 4C, 6-diamidino-2-phenylindole (Vector, Burlingame, CA, USA). The stained sections were analyzed using an LSM 700 confocal microscope (Zeiss, Gottingen, Germany). The z-stack confocal analysis was used to measure the colocalization of MAP2 with MCT2 and GFAP with MCT4 and to create Maximum Intensity Projection (MIP) and Ortho (2.5D) images for further colocalization clarification.

Statistical analyses were performed using the Statistical Package for Social Sciences software (version 25.0; IBM Corporation, Armonk, NY, USA). Levene's test was applied for the homogeneity of variance of the data. A two-way analysis of variance (ANOVA) test was used to examine the primary effects (Sex or Period or Housing) and interaction effects (Sex × Housing or Period × Housing), followed by Tukey's post-hoc test for comparisons within sex or period and between housing conditions under significant interaction. A two-way repeated measure ANOVA test was used to examine the primary and interaction effects within and between groups (5 × 4 factorial design) for the rotarod test. Post-hoc analysis was used to find where the significant differences were, and was identified at p < 0.01 using a Bonferroni adjustment as a multiple pairwise comparison. For the comparison of the OVX groups, a one-way ANOVA followed by Tukey's multiple comparison test was used. The data are expressed as the mean ± standard error of the mean (SEM), and all graphical artworks were produced using GraphPad Prism version 9.

For 8 weeks from 6 weeks of age, normal mice were exposed to either control cages or EE cages (Supplementary Figure 1A); the experimental scheme for normal models is shown in Figures 1A,C,D, and body weight change is noted in Figure 1B. Representative DXA images from each group are shown in Figure 1E. In DXA images, red dots indicate fat tissues, blue dots indicate non-fat tissues, and white dots indicate bone tissues. Significant housing effect was observed, indicating that fat reduction occurred after exposure to EE regardless of sex (Figure 1F). In blood biochemical analysis, significant gender effect and housing effect were noted in serum triglyceride (TG), indicating that EE mice have lower TG level than control mice regardless of sex, while female mice showed a higher TG level than male mice regardless of housing condition (Figure 1G). A significant housing effect in serum total cholesterol (TCHO) was also observed, indicating that exposure to EE can decrease TCHO level regardless of sex (Figure 1H). Metabolic cage analysis showed a significant housing effect in the respiratory exchange ratio, indicating that EE mice have higher RER ratio than control mice regardless of sex (Figure 1I) and significant sex effect on heat, indicating that female mice have higher heat than male mice regardless of housing condition (Figure 1J). Collectively, these data imply that female mice have different body composition compared to male mice, and long-term exposure to EE can induce fat reduction and change body composition, with resultant metabolic changes in both sexes of mice.

βHB enzyme-linked immunosorbent assay (ELISA) analysis indicated significant housing and sex effects on serum βHB levels (Figure 2A), indicating that female mice have higher βHB level than male mice regardless of housing, and EE mice have higher βHB level than control mice regardless of sex. Further analysis based on the estrus cycle indicated a trend on the interaction between housing and period (P = 0.0570). Significant housing effect and period effect on serum βHB levels in female mice were observed, indicating that serum βHB levels in proestrus and estrus (P/E) stage were higher than in metestrus and diestrus (M/D) stage regardless of housing, and EE mice have higher serum βHB levels than control mice regardless of period (Figure 2B). The stage of the estrus cycle was determined through vaginal cytology (Supplementary Figure 1B). Moreover, significant housing and sex effects were also indicated on βHB levels in the cerebral cortex and hippocampus, respectively (Figures 2C,D). This indicates that female mice have higher βHB levels than male mice regardless of housing, and EE mice have higher βHB levels than control mice regardless of sex. These data demonstrate that the baseline level of βHB is different between sexes, and EE can induce metabolic changes and alter body composition. The changes in female mice are due to the influence of the estrus cycle in response to the increased rate of lipolysis.

To examine the expression of βHB-related genes, qRT-PCR was conducted in the cerebral cortex (Figures 3A1–G1) and hippocampus (Figures 3A2–G2). There was a significant interaction between housing and gender in MCT2 [F_{(1, 44)} = 36.18, P = 0.0001], HDAC3 [F_{(1, 44)} = 10.96, P = 0.0019], and BDNF [F_{(1, 44)} = 20.85, P = 0.0001] in cerebral cortex. Moreover, there was a significant interaction between housing and gender in MCT2 [F_{(1, 44)} = 11.38, P = 0.0016], MCT4 [F_{(1, 44)} = 8.822, P = 0.0048], HDAC1 [F_{(1, 44)} = 42.80, P = 0.0001], HDAC2 [F_{(1, 44)} = 5.344, P = 0.0255], and BDNF [F_{(1, 44)} = 29.65, P = 0.0001] in hippocampus. Significant housing effects and sex effects are noted in the figures. Western blotting was conducted to examine the expression of βHB-related proteins in cerebral cortex (Figures 4A1–H1) and hippocampus (Figures 4A2–H2), and representative images are shown in Figures 4A1,A2, respectively. Significant housing or sex effect was noted in βHB-related proteins. Collectively, these data indicate that the expression of βHB-related genes and proteins was significantly regulated by EE and sex. In cerebral cortex and hippocampus, EE led to a significant increase in the level of βHB and a significant reduction in the level of HDAC, which consequently enhanced BDNF expression under the influence of sex.

To examine the uptake of βHB in cerebral cortex and hippocampus, histological assessments with GFAP, MCT4, MAP-2, and MCT2 were conducted. The representative images of GFAP and MCT4 in the cerebral cortex and hippocampus are shown in Figures 5A1,A2, respectively. The representative images of MAP-2 and MCT2 in the cerebral cortex and hippocampus are shown in Figures 5B1,B2, respectively. Significant sex and housing effect were noted both in the colocalization of GFAP and MCT4 in cerebral cortex (Figure 5C1) and the colocalization of MAP2 and MCT2 in hippocampus (Figure 5D1), indicating that female mice have the higher colocalization than male mice regardless of housing, and EE mice have the higher colocalization than control mice regardless of sex. Significant housing effect was noted in the colocalization of GFAP and MCT4 in cerebral cortex (Figure 5C2) and in the colocalization of MAP2 and MCT2 in hippocampus (Figure 5D2), implying that EE mice have higher colocalization than control mice regardless of sex. Raw intensity of GFAP and MAP2 for cerebral cortex and hippocampus are shown in Figures 5E1,F1,E2,F2. Significant housing effect was observed in raw MAP2 expression, demonstrating that EE mice showed a higher expression of MAP2 than control mice regardless of sex in cerebral cortex and hippocampus. Significant gender effect was observed in raw GFAP expression, indicating that female mice showed higher expression of GFAP than male mice regardless of housing in hippocampus. These combined data indicate that long-term exposure to EE can induce the higher βHB uptake as indicated by the higher colocalization rate in cerebral cortex and hippocampus by astrocytes and neurons.

To examine motor and cognitive function of normal mice, rotarod test and Y-maze test were conducted. In rotarod test, female EE mice significantly outperformed the other groups of mice at accelerating speed 4–80 rpm (Figure 6A), constant 48 rpm (Figure 6B), and constant 64 rpm (Figure 6C). Specifically, rotarod performance was significantly improved in EE mice from 2 to 8 weeks after EE treatment. In Y-maze test, significant housing effect was observed in the number of alterative behaviors (Figure 6D) and the number of entries (Figure 6E), demonstrating that the raw number of alternation and total entries were significantly decreased in EE mice compared to control mice regardless of gender. Significant housing effect was observed in the percentage of alternation, indicating that EE mice have higher percentage of alternation than control mice regardless of gender (Figure 6F). These data indicated that exposure to EE improves cognitive function. These combined data indicate that long-term exposure to EE improves motor and cognitive function.

To produce an estrogen-deficient model, OVX was conducted at 6 weeks of age (Supplementary Figure 1C), and the experimental scheme for OVX models is shown in Figure 7A. Body weight measurement for OVX group, OVX + E2 group, OVX + EE group, and OVX + E2/EE group is noted in Figure 7B, and the body weight was significantly decreased by exposure to EE and the synergistic effect of EE and E2. The level of 17β-estradiol was significantly increased by estrogen treatment and exposure to EE (Figure 7C). The representative DEXA images from each group are shown in Figure 7D, and significant fat reduction following estrogen treatment and exposure to EE was observed (Figure 7E). Moreover, in blood biochemical analysis, the synergistic effect of estrogen and exposure to EE was observed in serum TG (Figure 7F), decreasing the level of TG compared to OVX group but not in serum TCHO (Figure 7G). In indirect calorimetry, exposure to EE and estrogen treatment significantly increase RER (Figure 7H) and heat (Figure 7I) in OVX mice, indicating that both treatments can change body composition and induce metabolic changes. Collectively, exposure to EE and estrogen treatment can synergistically increase lipolysis, alter body composition, and metabolic changes in OVX mice.

To examine lipid accumulation in the liver and various brain regions of OVX mice, Oil Red O (ORO) staining, and H&E staining were conducted. The representative images of ORO-stained livers for each group are shown in Figure 8A, and the abnormal lipid accumulation was significantly decreased by the synergistic effect of EE and estrogen treatment (P = 0.0194, Figure 8B). Moreover, hepatic steatosis was observed in the OVX group and was diminished by exposure to EE and/or estrogen treatment (Figure 8C). The representative images of ORO-stained cerebral cortex, pia-mater, and subventricular zone (SVZ) are shown in Figures 8D–F, respectively. The quantification of ORO-stained cerebral cortex, pia-mater, and SVZ are shown in Figures 8G–I. Exposure to EE and estrogen treatment significantly reduced the abnormal accumulation of lipid droplets in cerebral cortex and SVZ. These data indicate that exposure to EE and estrogen treatment can induce lipolysis and reduce abnormal lipid accumulation in the liver and various brain regions.

βHB ELISA analysis indicated a significant increase in serum βHB after estrogen treatment and exposure to EE (Figure 9A). The synergistic effect of estrogen and EE was observed on the βHB level in the cerebral cortex and hippocampus (Figures 9B,C). In response to the increased rate of lipolysis, the level of βHB was significantly increased by EE and estrogen treatment in serum and the brain regions. Particularly, in female mice with EE and estrogen treatment, the level of βHB in cerebral cortex was significantly increased compared with OVX mice (P = 0.0319, Figure 9B), and the level of hippocampal βHB was significantly increased compared with OVX mice (P = 0.0014, Figure 9C).

To examine the expression of βHB-related genes, qRT-PCR was conducted in the cerebral cortex (Figures 10A1–G1) and hippocampus (Figures 10A2–G2). The significantly synergistic effect of EE and estrogen treatment was noted in MCT2, MCT4, HDAC1, HDAC2, HDAC3, and BDNF in both cerebral cortex and hippocampus. To examine the expression of βHB-related proteins, western blotting was conducted. The representative western blot images of βHB-related proteins in cerebral cortex and hippocampus are shown in Figures 11A1,A2, respectively. The significantly synergistic effect of EE and estrogen treatment was noted in MCT2, MCT4, HDAC2, HDAC3, and BDNF in cerebral cortex (Figures 11B1–H1). Moreover, the significantly synergistic effect of EE and estrogen treatment was noted in MCT2, MCT4, HDAC1, HDAC2, HDAC3, and BDNF in hippocampus (Figures 11B2–H2). Collectively, these data indicate that the expression of βHB-related genes and proteins was significantly regulated by exposure to EE and estrogen treatment. Particularly, the expressions of BDNF genes and proteins were significantly enhanced in the cerebral cortex (P = 0.0004, Figure 10G1; P = 0.0342, Figure 11H1) and hippocampus (P < 0.0001, Figure 10G2; P = 0.0025, Figure 11H2) of the female mice.

To examine the uptake of βHB in cerebral cortex and hippocampus, histological assessments with GFAP, MCT4, MAP-2, and MCT2 were conducted. The representative images of GFAP and MCT4 in the cerebral cortex and hippocampus are shown in Figures 12A1,A2, respectively. The representative images of MAP-2 and MCT2 in the cerebral cortex and hippocampus are shown in Figures 12B1,B2, respectively. The colocalization percent of GFAP with MCT4 (P < 0.0001, Figure 12C1) and MAP2 with MCT2 (P < 0.0001, Figure 12D1) in the cerebral cortex was significantly increased in OVX+E2/EE group compared to OVX group. In similar fashion, the colocalization percent of GFAP with MCT4 (P = 0.0012, Figure 12C2) and MAP2 with MCT2 (P < 0.0001, Figure 12D2) in the hippocampus was significantly increased in OVX + E2/EE group compared to OVX group (Figures 12C2,D2). Raw intensity of GFAP and MAP2 for cerebral cortex and hippocampus are shown in Figures 12E1,F1,E2,F2. These combined data indicate that long-term exposure to EE and estrogen treatment can synergistically induce the higher βHB uptake in cerebral cortex and hippocampus by astrocytes and neurons in OVX mice.

To examine motor, emotional, and cognitive function of OVX mice, hanging wire test, open-field test, and Y-maze test were conducted. Significant improvement in motor function following estrogen treatment and exposure to EE was observed in the hanging wire test (P < 0.0001, Figure 13A). Moreover, a significant reduction in anxiety level (P = 0.0036, Figure 13B) and cognitive improvement (P = 0.0001, Figure 13E) were observed in OVX mice after treatment with estrogen and EE. Raw number of alternative behaviors and number of entries in Y-maze are shown in Figures 13C,D, respectively. Using an OVX model, these data indicate that EE and estrogen treatment synergistically regulate the expression of βHB-related genes and proteins, thereby improving motor, cognitive, and emotional functions in female mice.