Wild-type C57BL/6J (C/EBPβ+/+) mice and C/EBPβ heterozygous knockout (C/EBPβ+/−) mice were purchased from the Jackson Laboratory (stock #000664 and #006873, respectively). Since some of the homozygous mutations are lethal on pure-strain backgrounds, C/EBPβ+/− mice were maintained as heterozygotes on a C57BL/6 strain background.

Thy1-human C/EBPβ mouse is a gift from Dr. Keqiang Ye. To generate Thy1-human C/EBPβ and ApoE4- human C/EBPβ transgenic mice of C57BL/6 J background, mouse genomic fragments containing homology arms (HAs) were amplified from a bacterial artificial chromosome (BAC) clone by using high-fidelity Taq, and were sequentially assembled into a targeting vector together with recombination sites and selection markers. After confirming the correctly targeted ES clones by Southern Blotting, we selected some clones for blastocyst microinjection, followed by founder production. Founders were confirmed as germline-transmitted via crossbreeding with wild-type mice. In the end, male F1 heterozygous mutant mice were confirmed as the final deliverables for this project.

The genotypes of transgenic mice were validated by polymerase chain reaction (PCR): Primer mix #1 (to distinguish Non-Tg and C/EBPβ-Tg): Forward: TGAAGCATTCCCTAATGAGCCAC, reverse: CTCGCCTCCTCCGGCCACTGCTAG. Primer mixes #2 (to distinguish Non-Tg, Tg and Tg/Tg): Forward1: AGAGTTGGTTGGTCCTCTCCT, reverse: GCCATTTAAGCCATGGGAAGTTAG, forward2: TGGACAGAGGAGCCATAACTGCAG.

Only male animals were used for the experiments. All mice were group-housed and kept under specific pathogen-free (SPF) conditions with a 12 h light/12 h dark cycle and with free access to food and water. Mice were randomly assigned to each group by using a random number table. The sample size was determined by Power and Precision (Biostat). All animal experimental protocols were approved by the Laboratory Animal Welfare Ethical Committee (IACUC) of Renmin Hospital of Wuhan University (IACUC Issue No. WDRM 20210102B).

Animal care and handling was performed according to the NIH animal care guidelines and Wuhan University guidelines. All procedures involving animals followed the ethical standards of Renmin Hospital of Wuhan University Institutional Animal Care and Use Committee.

At first, to investigate whether the HFD induced depressive behavior, we fed C57 BL/6 J wild-type mice with an HFD or chow diet for 2, 8, or 12 weeks and evaluated T2DM-associated phenotypes and multiple depression-related behaviors. Then, to find the key molecular factor underlying these phenotypes, we subjected the C57 BL/6 J wild-type mice fed with HFD for 12 weeks to a social interaction test (based on the subthreshold social defeat paradigm). We then classified the mice as susceptible (social interaction scores <100) or resilient (social interaction scores ≥100) subpopulation and used mice fed with a chow diet for 12 weeks as a control group. Next, to test the function of the key molecular factor C/EBPβ in HFD-induced depression-like behavior, we fed C/EBPβ+/−mice and C/EBPβ Tg mice with HFD or chow diet for 12 weeks and 2 weeks, respectively. Finally, to rescue depression-like behavior, C/EBPβ Tg mice were fed with an HFD or chow diet for 12 weeks and given treatment of anti-inflammatory drug aspirin, overexpression of BDNF in their hippocampus, and administration of small molecular BDNF mimetic compound at the eighth week. We only employ male mice in our study. All mice were 8–10 weeks old and had 1 week of adaptation to their surroundings before experiments.

Mice were fed with a chow diet or HFD (D12079B, Research Diets). We administered aspirin through their drinking water at pH 6.4 (120 mg/kg body weight/day) during the last 4 weeks on the HFD. The mice received vehicle or 7,8-DHF through their drinking water. To dissolve 7,8-DHF in water, 1 M NaOH was added drop wise to the water and stirred at room temperature overnight. The final concentration of 7,8-DHF was 22 mg/l (pH 7.6–7.8). Water (pH 7.6–7.8) was used as the vehicle control. Since the daily water intake of C57BL/6 J mice is about 7 ml/30 g body weight (Bachmanov et al., 2002), the oral dose of 7,8-DHF was ∼5 mg/kg/day. All animals are randomly assigned to the treatment or diet groups without considering any other variables. The protocol was reviewed and approved by Renmin Hospital of Wuhan University Institutional Animal Care and Use Committee.

The vectors of AAV8-human mature BDNF (AAV-BDNF) and AAV8-enhanced GFP (AAV-GFP) were generated, produced and purified by Virovek (Hayward, CA, United States; Jiao et al., 2016). AAV was injected stereotactically into mice under isoflurane anesthesia. We used the following coordinates for bilateral intracerebral injections: −1.5 mm anteroposterior, −2.06 mm mediolateral from the bregma, and −1.85 mm dorsoventral from the dual surface. Viral suspension (2 μl) containing 2 × 10⁹ vector genomes per μL was placed into each site at a rate of 0.25 μl min⁻¹ using a 10 μl glass syringe with a fixed needle. After the injection, we left the needle in place for 10 min and removed it slowly over 2 min. Mice were placed on a heating pad until they recovered from anesthesia. Four weeks after the stereotactic injection, we performed the multiple behavior test on the mice.

We assessed the susceptibility to stress of C57BL/6 J mice using the subthreshold social defeat stress test. This stress model resembles the chronic social defeat stress model, but C57BL/6 J mice are exposed to a new resident CD1 mouse and subjected to social defeat each day for 4 consecutive days rather than 10, and tested for social interaction on the 5^th day. In this subthreshold social defeat paradigm test, the defeated control mice do not show significant social avoidance and anhedonia in the subthreshold social defeat paradigm (Liu et al., 2012).

The tail suspension test (TST) reveals despair/depression-like behavior by recording the time suspended mice remain immobile. In brief, each mouse was individually suspended 20 cm above the floor with adhesive tape placed 1 cm from the tip of the tail. Animals were considered as immobile when they hung passively for 10 s without body movement. The time during which mice remained immobile was quantified in a total of 6 min. A longer immobility time than control mice indicated depression-like behaviors.

The forced swim test (FST) evaluates depression-like behaviors by measuring the time swimming mice spend immobile. Briefly, mice were individually placed into a glass cylinder (35 cm in height and 15 cm in diameter) filled with 10 cm of warm water (25 ± 1°C). After each trial, we renewed the water. We forced the mice to swim for 6 min, and recorded the immobility time during the final 4 min. Immobility was defined as floating or remaining motionless, without movement, except those motions necessary to keep the head above the water. The observers were blind to the treatment of the mice. An increased immobility time compared with control mice indicated depression-like behaviors.

The sucrose preference test (SPT) assesses anhedonia which is a key feature of depression. Anhedonia is defined as a percentage of sucrose preference below 65% (Strekalova et al., 2004; Scheggi et al., 2018). We used a 6-day sucrose preference protocol to examine the depression-like behaviors in wild-type, C/EBPβ+/− and Thy1-C/EBPβ Tg mice. In short, mice were first individually housed for a week, then treated with two bottles of normal water for 2 days, followed by two bottles of 2% sucrose solution for 2 days. After that, mice were water deprived for 24 h and then received access to two bottles, one filled with a 2% sucrose solution and the other with normal water for 2 h in the dark. The positions of the bottles were switched after 1 h. We then recorded the total consumption of each fluid and calculated the sucrose preference as the ratio of the consumption of sucrose solution to the consumption of both water and sucrose solution during the 2 h test (expressed as a percentage of sucrose intake versus total fluid intake). A lower sucrose preference than control mice indicated depression-like behavior.

To assess mobility and anxiety, we performed an open-field test (OF) in a rectangular chamber (50 × 50 × 50 cm) made of opaque white plastic. The ground area of the box was divided into two parts: a 35 × 35 cm central zone and the surrounding zone. Mice were gently put in the corner and allowed to explore the chamber for 5 min on two consecutive days. A video camera and a 25 W red light bulb were placed 180 cm above the center of the apparatus and recorded the movements with an ANY-maze video motility system. This system recorded the time spent in the central zone, the number of entries into the central zone, the total distance traveled in the whole field, and the average speed. Anxiety-related behavior was evaluated by recording these parameters on the second day. Lower center time and distance values than control mice indicated anxiety-related behavior.

We also evaluated anxiety through the elevated plus maze (EPM) test. Mice were tested for 5 min on an elevated plus maze apparatus on two consecutive days. The elevated plus maze, which was elevated 40 cm above the floor, consisted of a plus-shaped platform with two open arms (30 × 5 × 0.5 cm), two closed arms (30 × 5 × 15 cm), and a connecting central zone (5 × 5 cm). At the beginning of each test, mice were placed in the central zone, facing an open arm. The movements of the mice during a 5 min trial period were tracked by a video camera above the center of the maze and recorded with the ANY-maze software. To count as an entry into an arm, the center body of the mouse had to cross the border to the arm. We recorded and calculated the total distance traveled in the elevated plus maze apparatus, the number of entries into open or closed arms, and the time spent in open and closed arms. Next, we quantified mobility using the total distance traveled in the both arms and the total number of entries into any arm during the 5 min test period. Anxiety-like behavior was evaluated using the time spent in the open arms and the number of entries into the open arms on the second day. A shorter time spent in the open arms and longer time spent in the closed arms compared with the control mice indicated anxiety-related behaviors.

We assessed hippocampal-dependent memory by performing a novel object recognition (NOR) test in an open-field apparatus (50 × 50 × 50 cm). Before the test, mice were allowed to freely explore the testing apparatus for 10 min. On the first day, mice were presented with two identical objects (familiar objects). We placed the objects at the left and right corners of the area, and allowed the mice to freely explore the objects for 5 min. Then, on the second day, we replaced one of two familiar objects with another object (novel object), and again allowed the mice to freely explored for 5 min. The time spent in exploring each object (familiar and novel objects) was recorded using a digital video camera and recorded with ANY-maze software. We defined “exploration” as touching the object (except with the tail) or sniffing the object (distance <2 cm). To analyze the recognition ability, we defined a discrimination index as the ratio of the time spent exploring the novel object over the time spent exploring the familiar and novel objects. A preference for the novel object indicated intact spatial recognition memory. Mice that did not explore any of the two objects during the pre-test were excluded from the analysis.

Bilateral hippocampal tissues were grinded and lysed on ice for 30 min with 80 μl of ice-cold lysis buffer containing protease and phosphatase inhibitors. We then centrifuged the samples at maximum speed for 15 min at 4°C, and collected the supernatant, and quantified proteins with a Thermo BCA Protein Assay Kit (Cat#23227). Protein samples were boiled in SDS loading buffer for 10 min, and then were separated by 8–12% SDS-PAGE, and transferred to nitrocellulose membranes. The membranes were blocked with 5% skim milk in Tris-buffered saline containing 0.1% Tween-20 (TBS-T) for 1 h at room temperature and incubated overnight at 4°C with the appropriate primary antibodies. After washed 4–6 times with TBS-T, the membrane was incubated for 1 h at room temperature with horseradish peroxidase (HRP)-conjugated anti-mouse secondary antibodies (1:5000; BL001A, Biosharp Life Sciences, China) and anti-rabbit secondary antibodies (1:5000; BL003A, Biosharp Life Sciences, China). Then, after washed 6–8 times with TBS-T, the membrane was immersed in enhanced chemiluminescence reagents, and images were captured by ChemiDoc™ Touch Imaging System. ImageJ software was used to analyze the bands. β-actin or tubulin was used as the loading control and all the experiments were performed at least three times.

Surface receptor cross-linking with bis(sulfosuccinimidyl)suberate (BS³) was performed as described previously (Lu et al., 2014). We isolated the brains, cut coronal hippocampal slices (300 μm), placed them into small tubes containing ice-cold artificial cerebrosoinal fluid, and immediately added 2 mM BS³ (21,580, Thermo Scientific, Rockford, USA). Then the tissue was cross-linked with gentle agitation (30 min at 4°C) and this reaction was stopped by adding 100 mM glycine (15 min at 4°C). Next, we centrifuged the tissue, resuspended it in ice-cold lysis buffer (50 mM Tris–HCl, 100 mM NaCl, 1% Nonidet P-40, 10 mM EDTA, 20 mM NaF, 1 mM PMSF, 3 mM Na₃VO₄, and protease inhibitor mixture), and centrifuged it at 12000 g for 15 min at 4°C. Finally, the supernatant fraction was aliquoted and stored at −80°C before use in the Western blotting.

Total RNA was isolated from tissues using a FastPure Cell/Tissue Total RNA Isolation Kit (cat. RC112-01, Vazyme, China) according to the manufacturer’s instructions and was quantified by the Nanodrop apparatus. After that, 1 μg of total RNA was used for cDNA synthesis with an NG Script I cDNA Synthesis kit (NG047S, HLINGENE, China) according to the manufacturer’s instructions. Quantitative real-time PCR (qRT-PCR) was performed in a Bio-rad Real-Time PCR System. For qRT-PCR, samples were heated to 95°C for 10 min followed by 40 cycles of 95°C for 15 s, 60°C for 30 s, and 72°C for 30 s. PCRs for CEBPB and BDNF were performed in triplicate. The relative quantification of gene expression was normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA levels and calculated using the ΔΔCt method. Gene expression analyses were expressed as mRNA levels relative to controls. Real-time PCR probes were bought from TaqMan® (Thermo Fisher Scientific).

Mice abdominal aortic blood samples were collected and centrifuged (2,000 rpm at 4°C for 10 min) and then the supernatant was collected and stored at −80°C until use. Quantitative determination of IL-1β, IL-6, and TNF-α were performed using commercially available high-sensitivity kits according to the manufacturer’s instructions.

Peripheral blood mononuclear cells (PBMC) were isolated from fresh blood by density gradient centrifugation over Lympholyte® Mammal (Haoyang Biological Manufacture Co. Ltd., Tianjin, China) according to the manufacturer’s instructions. Briefly, fresh blood was diluted in sterile diluent with a ratio of 1:1, layered onto the Lymphoprep solution, and centrifuged at 450 g at room temperature for 20 min. After that, the PBMC layer at the interface was carefully collected and transferred into a 15-mL tube, which washed twice with sterile phosphate-buffered saline (PBS, pH 7.4) and then centrifuged at 250 g for 10 min. The harvested cell pellet was stored at −80°C until analysis.

We used free-floating 20 μm brain sections for immunofluorescence staining. The brain sections were washed three times with PBS and blocked with 1% bovine serum albumin (w/v) and 0.3% Triton X-100 (v/v) for 30 min, and were incubated with primary antibodies at 4°C overnight. On the second day, after washed three times with PBS, the brain sections were incubated at room temperature with a mixture of labeled secondary antibodies for 2 h. Finally, DAPI (1:1,000; Sigma-Aldrich) was used for staining nuclei for 5 min followed by washed three times with PBS. Images were acquired with a Leica Confocal Imaging System.

Mice brains were fixed in 10% formalin (v/v) for 24 h and then immersed in 3% potassium bichromate (w/v) for 3 days in the dark. The solution was changed every day. The brains were transferred into a 2% silver nitrate (w/v) solution and incubated for 7 days in the dark. Vibratome sections were cut at 50 μm, air dried for 10 min, dehydrated through 95 and 100% ethanol, cleared in xylene and coverslipped. Spine numbers were counted in Image J software.

Mice were anaesthetized with isoflurane, decapitated, and their brains dropped in ice-cold a-CSF containing 124 mM NaCl, 3 mM KCl, 1.25 mM NaH₂PO₄, 6.0 mM MgCl₂, 26 mM NaHCO₃, 2.0 mM CaCl₂, and 10 mM glucose. Hippocampi were dissected and cut into 400-mm thick transverse slices with a vibratome. After incubation at room temperature (23–24°C) in a-CSF for 60–90 min, slices were placed in a recording chamber (RC-22C, Warner Instruments) on the stage of an up-right microscope (Olympus CX-31) and perfused at a rate of 3 ml/min with a-CSF (containing 1 mM MgCl₂) at 23–24°C. A 0.1 MU tungsten monopolar electrode was used to stimulate the Schaffer collaterals. The field excitatory post-synaptic potentials (fEPSPs) were recorded in the CA1 stratum radiatum by a glass microelectrode filled with a-CSF with a resistance of 3–4 MU. The stimulation output (Master-8; AMPI, Jerusalem) was controlled by the trigger function of an EPC9 amplifier (HEKA Elektronik, Lambrecht, Germany). fEPSPs were recorded under current-clamp mode. Data were filtered at 3 kHz and digitized at sampling rates of 20 kHz using Pulse software (HEKA Elektronik). The stimulus intensity (0.1 ms duration, 10–30 mA) was set to evoke 40% of the maximum fEPSP and the test pulse was applied at a rate of 0.033 Hz. The long-term potentiation (LTP) of fEPSPs was induced with three theta-burst-stimulation (four pulses at 100 Hz, repeated three times with a 200-ms interval). The magnitudes of LTP are expressed as the mean percentage of the baseline fEPSP initial slope.

Mouse brain tissue samples were lysed in lysis buffer (50 mM Tris, pH 7.4, 40 mM NaCl, 1 mM EDTA, 0.5% Triton X-100, 1.5 mM Na3VO4, 50 mM NaF, 10 mM sodium pyrophosphate, 10 mM sodium-glycerophosphate, supplemented with protease inhibitors cocktail) on ice for 30 min. For the in vitro experiment, cells were lysed in RIPA buffer (pH 7.5, 20 mM Tris–HCl, 150 mM NaCl, 1 mM Na₂EDTA, 1 mM EGTA, 1% Triton X-100, 2.5 mM sodium pyrophosphate, 1 mM beta-glycerophosphate, 1 mM Na₃VO₄, 1 μg/ml leupeptin, 1 mM phenylmethylsulfonyl fluoride) on ice for 30 min. The lysates were centrifuged for 10 min at 15,000 rpm. We quantified proteins in the supernatant using a Coomassie Brilliant Blue protein assay kit (Bio-Rad). Next, we boiled the supernatant mixed with the same amount of SDS loading buffer. After SDS- PAGE, the samples were transferred to a nitrocellulose membrane. The membranes were blocked in 5% non-fat milk for 1 h at room temperature, then incubated with primary antibody at 4°C overnight. Then the blots were incubated with IRDye 800CW-conjugated affinity-purified anti-mouse or anti-rabbit IgG secondary antibody (Rockland). Immunoreactive bands were visualized using an Odyssey Infrared Imaging System (Licor Biosciences, Lincoln, NE, United States).

We performed a glucose tolerance test (GTT) on mice fasted for 16 h after a peritoneal injection of d-glucose (2 g/kg body weight). Each mouse was weighed before the injection to determine the appropriate dose. Blood samples from the tail vein were collected at 0, 15, 30, 60, and 90 min post-injection.

For the insulin tolerance test (ITT), the mice fasted for 6 h. Each mouse was weighted and administered a peritoneal insulin injection of insulin (0.75 U/kg body weight). Blood glucose level from the tail vein were collected at 0, 15, 30, 60, and 90 min post-injection (Vinué and González-Navarro, 2015).

We analyzed all data using GraphPad Prism (GraphPad Software) and expressed the results as the mean ± standard error of the mean (SEM) from three or more independent experiments. All the statistical procedures can be found in the figure legends for each experiment, including the statistical tests used, number of mice used. Sample sizes were determined by Power and Precision (Biostat). We compared group pairs using an unpaired t-test with Welch’s correction. For multiple-group comparisons, one-way ANOVA and Bonferroni’s multiple comparison test was applied. The two-way ANOVA and Bonferroni’s post hoc test compared the differences between groups that have been split on two independent factors. Values of p < 0.05 indicated statistical significance.

The data that support the findings of this study are available on request from the corresponding author.

To confirm whether the HFD consumption induced depressive behavior, we first confirmed that we had successfully established an HFD-induced insulin resistance murine model. Next, we evaluated multiple depression-related behaviors in mice fed with an HFD or chow diet for 2, 8, or 12 weeks (Figure 1A). The body weight curve showed that HFD-fed mice gained more body weight than chow diet mice (Supplementary Figure S2A). The ITT at three time points (30, 60, and 90 min) revealed that mice fed with an HFD for 12 weeks displayed more insulin intolerance than chow diet mice (Supplementary Figure S2B). Additionally, the GTT indicated that chow diet mice demonstrated much more glucose tolerance than those fed with an HFD for 12 weeks, albeit only at the 60 min time point (Supplementary Figure S2C). Impaired insulin tolerance is demonstrated for three time points (30, 60, and 90 min) after insulin administration. However, in the glucose tolerance test, only 60 min time point appears to be changed between the chow and HFD groups. Environmental stress experienced by mice in the early stages of GTT may account for the differences in ITT and GTT results. Nevertheless, these results indicated that the HFD for 12 consecutive weeks successfully induced insulin resistance in mice. In behavioral tests, we found that 8 weeks of HFD induced depression-like behaviors (such as increased immobility time in the TST and FST), and this effect was more pronounced in mice fed an HFD for 12 weeks (Chow-2 67.72 ± 7.565 vs. HFD-2 80.96 ± 7.187, p > 0.05; Chow-8 61.91 ± 6.334 vs. HFD-8112.2 ± 11.1, p = 0.0026; Chow-12 74.85 ± 7.588 vs. HFD-12132.3 ± 12.86, p = 0.0003; Figure 1B), (Chow-2 78.16 ± 4.693 vs. HFD-2 86.8 ± 7.369, p > 0.05; Chow-8 75.75 ± 8.184 vs. HFD-8125.2 ± 10.43, p = 0.0036; Chow-12 80.12 ± 6.851 vs. HFD-12138.8 ± 14.26, p = 0.0003; Figure 1C). During the pre-treatment for the SPT, we observed no significant differences in the baseline sucrose consumption among the different groups (Figure 1D). After 8- or 12-weeks chow diet, mice displayed a higher sucrose preference than HFD mice, indicating that HFD induced depression-like behaviors (Chow-2 73.58 ± 3.733 vs. HFD-2 68.85 ± 3.914, p > 0.05; Chow-8 71.34 ± 4.305 vs. HFD-8 60.28 ± 4.346, p > 0.05; Chow-12 67.91 ± 4.883 vs. HFD-12 45.99 ± 3.365, p = 0.0045; Figure 1E).

To test whether these mice also exhibited anxiety-related behaviors, we performed open field (OF) test and elevated plus maze (EPM) test. In the OF test, the 8-week HFD group spent less time in the center and traveled shorter distances than the chow diet group, and this effect was more prominent in the 12-week HFD group (Chow-2 8.164 ± 0.6582 vs. HFD-2 7.692 ± 0.7131, p > 0.05; Chow-8 8.929 ± 0.629 vs. HFD-8 5.213 ± 0.4137, p = 0.0005; Chow-12 8.607 ± 0.7214 vs. HFD-12 4.06 ± 0.3762, p < 0.0001; Figure 1F, Chow-2775.8 ± 63.97 vs. HFD-2792.3 ± 80.04, p > 0.05; Chow-8800.1 ± 98.47 vs. HFD-8558.7 ± 64.43, p > 0.05; Chow-12721.1 ± 78.87 vs. HFD-12406 ± 37.62, p = 0.0449; Figure 1G). In the EPM test, the 8-and 12-week HFD groups spent significantly less time in the open arm and more time in the closed arm than the control group (Chow-2 52.88 ± 3.113 vs. HFD-2 51.81 ± 5.522, p > 0.05; Chow-8 56.08 ± 5.014 vs. HFD-8 40.07 ± 2.938, p > 0.05; Chow-12 54.5 ± 4.403 vs. HFD-12 34.25 ± 2.558, p = 0.0105; Figure 1H), (Chow-2167.1 ± 3.113 vs. HFD-2168.2 ± 5.522, p > 0.05; Chow-8157.9 ± 5.014 vs. HFD-8175.2 ± 3.657, p > 0.05; Chow-12165.5 ± 4.403 vs. HFD-12185.8 ± 2.558, p = 0.0137; Figure 1I). These results indicate that the HFD induced anxiety-related behaviors in mice.

To determine whether HFD impaired memory, we performed a novel object recognition (NOR) test. We found that the 12-week HFD group spent less time exploring new objects than familiar ones. Meanwhile, there was no difference among groups in the pre-test, indicating that 12-week HFD mice showed memory impairment (Figure 1J; Chow-2 0.7426 ± 0.04638 vs. HFD-2 0.7360 ± 0.04873, p > 0.05; Chow-8 0.7060 ± 0.06099 vs. HFD-8 0.6042 ± 0.6099, p > 0.05; Chow-12 0.6978 ± 0.05283 vs. HFD-12 0.4659 ± 0.05434, p = 0.0399; Figure 1K).

Additionally, to confirm that the changes in mobility observed in the OF test and FST were due to anhedonia and not to a large weight gain, we chose 52-week-old wild-type mice fed with a chow diet as an additional control. These mice have similar body weights as young HFD mice, but have no demonstrated insulin resistance (Supplementary Figures S2E–G; WT-Chow-20–22 weeks of age 35.38 ± 1.580 vs. WT-HFD-20–22 weeks of age 43.19 ± 2.593, p = 0.0473; WT-Chow-20–22 weeks of age 35.38 ± 1.580 vs. WT-Chow-52 weeks of age 44.75 ± 2.823, p = 0.0302; Supplementary Figure S2E). In addition, the 52-week-old wild-type mice showed normal mobility in open-field test and FST compared with wild-type-HFD mice-20–22 weeks of age (Supplementary Figures S2H–J; WT-Chow-20-22 weeks of age 8.607 ± 0.7214 vs. WT-HFD-20–22 weeks of age 4.060 ± 0.3762, p < 0.001; Supplementary Figure S2H; WT-Chow-20–22 weeks of age 721.1 ± 78.87 vs. WT-HFD-20–22 weeks of age 418.5 ± 30.87, p = 0.0029; WT-HFD-20-22 weeks of age 418.5 ± 30.87 vs. WT-Chow-52 weeks of age 755.6 ± 76.41, p = 0.0037; Supplementary Figure S2I; WT-Chow-20-22 weeks of age 80.12 ± 6.851 vs. WT-HFD-20-22 weeks of age 138.8 ± 14.26, p < 0.001; WT-HFD-20–22 weeks of age 138.8 ± 14.26 vs. WT-Chow-52 weeks of age 76.03 ± 3.505, p < 0.001; WT-HFD-20-22 weeks of age 4.060 ± 0.3762 vs. WT-Chow-52 weeks of age 9.012 ± 0.6173, p < 0.001; Supplementary Figure S2J). These results indicated that HFD caused depression-like behaviors in wild-type mice in a time-dependent manner.

To investigate whether neuroinflammation-activated C/EBPβ plays a role in depression, we first fed WT mice with HFD for 12 weeks and then classified them based on the subthreshold social defeat paradigm. This subthreshold social defeat stress itself does not cause significant social avoidance and anhedonia in mice model (Liu et al., 2012). After exposure to subthreshold social defeat stress, we subjected these mice to the social interaction test and determined their social interaction scores (Supplementary Figures S3A,B). Next, we sorted the mice as susceptible (scores <100) or resilient (scores ≥100) subpopulation. Furthermore, mice fed with a chow diet for 12 weeks served as control group (Control 151.3 ± 7.587 vs. Susceptible 52.66 ± 3.873, p < 0.0001; Susceptible 52.66 ± 3.873 vs. Resilient 160 ± 7.928, p < 0.0001; Figure 2A), (Supplementary Figure S3B). We found that, compared to control and resilient group, susceptible mice showed a significantly lower sucrose preference (Control 65.04 ± 6.78 vs. Susceptible 28.99 ± 2.304, p < 0.0001; Susceptible 28.99 ± 2.304 vs. Resilient 74.74 ± 4.322, p < 0.0001; Figure 2B), which was far below 50%, indicating these susceptible animals were actually anhedonia. Anhedonia is defined as a percentage of sucrose preference below 65% (Strekalova et al., 2004; Scheggi et al., 2018). Moreover, susceptible mice had dramatically higher levels of pro-inflammatory cytokines (IL-1β, IL-6, and TNF-α, tested by ELISA) in their brain than control and resilient mice. [IL-1β: Control 1 ± 0.1067 vs. Susceptible 3.084 ± 0.2909, p < 0.0001; Susceptible 3.084 ± 0.2909 vs. Resilient 1.237 ± 0.1544, p < 0.0001; Figure 2C (left); IL-6: Control 1 ± 0.2326 vs. Susceptible 3.21 ± 0.2411, p = 0.0002; Susceptible 3.21 ± 0.2411 vs. Resilient 1.408 ± 0.3625, p = 0.0013; Figure 2C (middle); TNF-α: Control 1 ± 0.1342 vs. Susceptible 1.446 ± 0.1289, p = 0.0318; Susceptible 1.446 ± 0.1289 vs. Resilient 0.9 ± 0.09961, p = 0.0191; Figure 2C (right)]. Consistent with these findings, Iba-1 and GFAP immunostaining indicated that susceptible mice had higher microglia and astrocyte counts in the hippocampus than control and resilient mice (Figure 2D, Iba-1: Control 1.012 ± 0.1864 vs. Susceptible 4.361 ± 0.3404, p < 0.0001; Susceptible 4.361 ± 0.3404 vs. Resilient 1.55 ± 0.1827, p < 0.0001; GFAP: Control 1.017 ± 0.1474 vs. Susceptible 5.984 ± 0.3781, p < 0.0001; Susceptible 5.984 ± 0.3781 vs. Resilient 1.652 ± 0.3732, p < 0.0001; Figure 2E).

Since C/EBPβ is an inflammation associated transcription factor, we analyzed C/EBPβ mRNA levels in peripheral blood mononuclear cells and hippocampus. We found that susceptible mice had markedly higher C/EBPβ mRNA levels than control mice (Control 1 ± 0.12 vs. Susceptible 5.588 ± 0.8268, p < 0.0001; Susceptible 5.588 ± 0.8268 vs. Resilient 1.452 ± 0.2283, p < 0.0001; Figure 2F), (Control 1 ± 0.2072 vs. Susceptible 3.846 ± 0.637, p = 0.0009; Susceptible 3.846 ± 0.637 vs. Resilient 1.615 ± 0.3232, p = 0.0068; Figure 2G), Additionally, the hippocampal C/EBPβ mRNA levels were negatively correlated with social interaction ratio (Figure 2H), suggesting that C/EBPβ plays a critical role in promoting HFD-triggered depression-like behaviors in WT mice. We also quantified BDNF mRNA in hippocampus and found that they followed a trend opposite to that of the C/EBPβ levels (Control 1 ± 0.1638 vs. Susceptible 0.4285 ± 0.1027, p = 0.0185; Susceptible 0.4285 ± 0.1027 vs. Resilient 0.7464 ± 0.1045, p = 0. 0486; Figure 2I).

To further confirm the role of C/EBPβ in depression, we conducted western blotting and immunostaining experiments. We discovered that susceptible mice had notably higher total C/EBPβ and phospho-C/EBPβ Thr188 (pC/EBPβ) levels than control and resilient mice, indicating the activation of C/EBPβ in the hippocampus [Figure 2J; pC/EBPβ/C/EBPβ: Control 1 ± 0.06014 vs. Susceptible 2.385 ± 0.2307, p = 0.0002; Susceptible 2.385 ± 0.2307 vs. Resilient 1.215 ± 0.1891, p = 0.0008; Figure 2K (left); C/EBPβ/Tubulin: Control 1 ± 0.1437 vs. Susceptible 1.812 ± 0.2172, p = 0.0273; Susceptible 1.812 ± 0.2172 vs. Resilient 1.112 ± 0.2064, p = 0.0387; Figure 2K (middle); BDNF/ Tubulin: Control 1 ± 0.123 vs. Susceptible 0.5437 ± 0.1019, p = 0.023; Susceptible 0.5437 ± 0.1019 vs. Resilient 0.9238 ± 0.08653, p = 0.0462; Figure 2K (right)]. Conversely, Western blotting and ELISA experiments revealed that susceptible mice had lower hippocampal BDNF protein levels than control and resilient mice (Control 1 ± 0.07586 vs. Susceptible 0.4459 ± 0.09788, p = 0.0009; Susceptible 0.4459 ± 0.09788 vs. Resilient 0.8333 ± 0.0752, p = 0.0153; Figure 2N). In addition, there was a higher density of pC/EBPβ immunofluorescence in susceptible mice hippocampus than that in the control group and pC/EBPβ was mostly located in the neurons, which are NeuN positive cells [Figure 2L; Control 0.9592 ± 0.096 vs. Susceptible 2.064 ± 0.2787, p = 0.0044; Susceptible 2.064 ± 0.2787 vs. Resilient 1.26 ± 0.2425, p = 0.047; Figure 2M (left); Control 1.01 ± 0.05281 vs. Susceptible 6.555 ± 0.5183, p < 0.0001; Susceptible 6.555 ± 0.5183 vs. Resilient 1.694 ± 0.2665, p < 0.0001; Figure 2M (right)]. Thus, we confirmed that HFD-induced neuroinflammation activated C/EBPβ in hippocampal neurons and further downregulated BDNF in anhedonia mice.

To assess the impact of C/EBPβ on HFD-elicited depression-like behaviors, we bred WT(C/EBPβ+/+) and C/EBPβ heterozygous knockout (C/EBPβ+/−) mice and treated them with HFD or chow diet for 12 weeks, then evaluated a series of depression-related behaviors and molecular component levels (Figure 3A). Since some of C/EBPβ homozygous knockout (C/EBPβ−/−) mice had a lethal embryonic phenotype and were difficult to breed, we focused on heterozygous knockout mice in this study. The genotyping of heterozygous knockout mice was confirmed by genomic PCR (Supplementary Figure S4A). Quantitative RT-PCR (qRT-PCR) revealed that the mRNA level of C/EBPβ in hippocampus was highly increased in WT-HFD mice compared to WT-chow diet mice, while there was no significant difference between HFD-fed and chow diet-fed C/EBPβ+/− mice (WT-Chow 1 ± 0.1291 vs. WT-HFD 2.4 ± 0.246, p < 0.0001; C/EBPβ+/−-Chow 0.5234 ± 0.04809 vs. C/EBPβ+/−-HFD 0.5377 ± 0.08152, p > 0.05; Figure 3B). However, its downstream target BDNF showed the opposite trend, with lower levels in WT mice fed with HFD than those fed with a chow diet. However, the mRNA level was largely upregulated in C/EBPβ+/− mice than that in WT mice even though there was no significant change in C/EBPβ+/− mice either fed with HFD or chow diet. (WT-Chow 1 ± 0.1279 vs. WT-HFD 0.5502 ± 0.1038, p = 0. 0463; C/EBPβ+/−-Chow 1.224 ± 0.1883 vs. C/EBPβ+/−-HFD 1.154 ± 0.1361, p > 0.05; Figure 3C).

Next, we quantified the hippocampal levels of the pro-inflammatory cytokines including TNF-α, IL-1β and IL-6 in mice hippocampus by ELISA and found that they were all significantly increased in WT-HFD mice versus WT-chow diet mice while there was no significant change in C/EBPβ+/− mice either fed with HFD or chow diet. [TNF-α: WT-Chow 1 ± 0.1301 vs. WT-HFD 1.59 ± 0.2418, p = 0. 0407; C/EBPβ+/−-Chow 1.184 ± 0.2467 vs. C/EBPβ+/−-HFD 1.347 ± 0.1695, p > 0.05; Figure 3D (left); IL-1β: WT-Chow 1.007 ± 0.1532 vs. WT-HFD 2.475 ± 0.4057, p = 0. 0029; C/EBPβ+/−-Chow 1.08 ± 0.196 vs. C/EBPβ+/−-HFD 1.554 ± 0.1508, p > 0.05; Figure 3D (middle); IL-6: WT-Chow 1 ± 0.1022 vs. WT-HFD 2.89 ± 0.2625, p < 0.0001; C/EBPβ+/−-Chow 0.8333 ± 0.1604 vs. C/EBPβ+/−-HFD 1.387 ± 0.3051, p > 0.05; Figure 3D (right)]. The ELISA and western blotting analysis indicated that pC/EBPβ and C/EBPβ were upregulated in WT-HFD mice compared to WT-chow diet mice, while its downstream target BDNF showed the opposite tendency (Figures 3E–G; WT-Chow 1 ± 0.04092 vs. WT-HFD 0.66 ± 0.06057, p = 0. 0091; C/EBPβ+/−-Chow 1.096 ± 0.0941 vs. C/EBPβ+/−-HFD 0.9267 ± 0.1983, p > 0.05; Figure 3E; pC/EBPβ/C/EBPβ: WT-Chow 1 ± 0.0733 vs. WT-HFD 1.602 ± 0.2543, p = 0.0265; C/EBPβ/Tubulin: WT-Chow 1 ± 0.03988 vs. WT-HFD 1.878 ± 0.1962, p < 0.0001; BDNF/Tubulin: WT-Chow 1 ± 0.05771 vs. WT-HFD 0.6226 ± 0.06939, p = 0.0029; Figure 3G). However, there was no significant difference in C/EBPβ+/− mice either fed with HFD or chow diet. (Figure 3F; pC/EBPβ/C/EBPβ: C/EBPβ+/−-Chow 0.861 ± 0.105 vs. C/EBPβ+/−-HFD 1.198 ± 0.86, p > 0.05; C/EBPβ/Tubulin: C/EBPβ+/−-Chow 0.5835 ± 0.04955 vs. C/EBPβ+/−-HFD 0.6714 ± 0.07576, p > 0.05; BDNF/Tubulin: C/EBPβ+/−-Chow 1.186 ± 0.07754 vs. C/EBPβ+/−-HFD 1.051 ± 0.1829, p > 0.05; Figure 3G).

In alignment with these discoveries, we found that depression-related behaviors induced by HFD were more obvious in WT-HFD mice versus WT-chow diet mice both in TST and FST (WT-Chow 62.27 ± 14.26 vs. WT-HFD 159.8 ± 20.61, p < 0.0001; C/EBPβ+/−-Chow 63.78 ± 7.404 vs. C/EBPβ+/−-HFD 74.31 ± 8.298, p > 0.05; Figure 3H; WT-Chow 74.19 ± 7.996 vs. WT-HFD 151.2 ± 15.47, p < 0.0001; C/EBPβ+/−-Chow 67.35 ± 7.515 vs. C/EBPβ+/−-HFD 75.64 ± 7.691, p > 0.05; Figure 3I). Whereas chow-diet mice displayed sucrose preference, HFD mice, especially WT-HFD mice showed less sucrose consumption (WT-Chow 71.71 ± 6.688 vs. WT-HFD 35.38 ± 4.611, p = 0.0008; C/EBPβ+/−-Chow 73.3 ± 6.928 vs. C/EBPβ+/−-HFD 71.35 ± 5.48, p > 0.05; Figure 3K). However, we found no significant difference among the different groups during the SPT pre-treatment (Figure 3J). Remarkably, the loss of C/EBPβ ameliorated the above behavioral dysfunctions in C/EBPβ+/− mice (Figures 3H–K). Hence, knocking down of C/EBPβ alleviates HFD-induced depression-like behaviors.

To further examine the effect of C/EBPβ in depression, we treated neuronal human C/EBPβ transgenic mice (Thy1-C/EBPβ Tg mice) and their littermate WT mice with HFD or chow diet for only 2 weeks before performing depression associated behaviors and molecular tests (Figure 4A). The neuronal-specific expression of human C/EBPβ transgene was driven by Thy1 promoter (Supplementary Figure S4B), and we validated C/EBPβ transgenic mice by genotyping strategy (Supplementary Figures S4C,D). Neuronal-specific expression of human C/EBPβ was validated in the recent reports (Ahn et al., 2021b; Wang et al., 2022). A Quantitative RT-PCR experiment showed that HFD-fed C/EBPβ transgenic mice had higher hippocampal human C/EBPβ mRNA levels than chow diet-fed C/EBPβ transgenic mice. Neither HFD-fed nor chow diet-fed wild-type mice expressed detectable levels of human C/EBPβ mRNA (C/EBPβ Tg-Chow 1 ± 0.06741 vs. C/EBPβ Tg-HFD 1.504 ± 0.2129, p = 0.0246; Figure 4B). Correspondingly, Thy1-C/EBPβ transgenic HFD mice had markedly lower levels of BDNF than that in the other groups (WT-Chow 1 ± 0.1022 vs. WT-HFD 1.188 ± 0.1561, p > 0.05; C/EBPβ Tg-Chow 0.7064 ± 0.09482 vs. C/EBPβ Tg-HFD 0.4294 ± 0.06923, p = 0. 036; Figure 4C).

Next, we quantified pro-inflammatory cytokines (TNF-α, IL-1β and IL-6) in mice hippocampus and found that HFD C/EBPβ transgenic mice had notably higher IL-1β and IL-6 levels than their corresponding-chow diet mice while there was no significant change between WT mice groups after a short-time HFD treatment. Meanwhile, TNF-α was not significantly increased between mice fed with HFD or chow diet TNF-α was not significantly increased between corresponding diet mice but there was an increasing trend between WT-chow and C/EBPβ Tg-HFD mice (TNF-α: WT-Chow 1 ± 0.1022 vs. C/EBPβ Tg-HFD 1.498 ± 0.1881, p = 0. 0478; IL-1β: WT-Chow 1 ± 0.136 vs. WT-HFD 1.045 ± 0.01989, p > 0.05; C/EBPβ Tg-Chow 1.35 ± 0.1382 vs. C/EBPβ Tg-HFD 3.113 ± 0.6486, p = 0.0089; IL-6: WT-Chow 1 ± 0.1724 vs. WT-HFD 1.238 ± 0.0274, p > 0.05; C/EBPβ Tg-Chow 1.787 ± 0.3637 vs. C/EBPβ Tg-HFD 4.293 ± 0.6079, p = 0.0006; Figure 4D). Western blotting analysis indicated that human C/EBPβ levels were higher in C/EBPβ transgenic-HFD mice than that in C/EBPβ transgenic-chow diet mice (Figures 4E–G; C/EBPβ Tg-Chow 1 ± 0.05592 vs. C/EBPβ Tg-HFD 2.052 ± 0.3112, p = 0.0018; Figure 4G). Its downstream target BDNF showed the opposite tendency both in ELISA and in western blotting (WT-Chow 1 ± 0.0623 vs. WT-HFD 1.053 ± 0.159, p > 0.05; C/EBPβ Tg-Chow 0.8163 ± 0.03716 vs. C/EBPβ Tg-HFD 0.5484 ± 0.04668, p = 0.0049; Figures 4E,F), (WT-Chow 1 ± 0.2048 vs. WT-HFD 1.068 ± 0.1424, p > 0.05; C/EBPβ Tg-Chow 0.7601 ± 0.1085 vs. C/EBPβ Tg-HFD 0.4026 ± 0.1044, p = 0.0383; Figure 4G).

Consistent with these molecular level findings, the TST and FST results showed that HFD-induced depression-like behaviors were more remarkable in C/EBPβ transgenic HFD mice than that in the corresponding chow diet mice and in both wild-type groups (WT-Chow 76.5 ± 7.06 vs. WT-HFD 84.53 ± 4.825, p > 0.05; C/EBPβ Tg-Chow 94.9 ± 10.62 vs. C/EBPβ Tg-HFD 143.9 ± 14.5, p = 0.008; Figure 4H; WT-Chow 76.81 ± 6.882 vs. WT-HFD 63.41 ± 6.776, p > 0.05; C/EBPβ Tg-Chow 82.77 ± 8.465 vs. C/EBPβ Tg-HFD 135.1 ± 15.08, p = 0.0038; Figure 4I). In the SPT, C/EBPβ transgenic HFD mice consumed less sucrose than chow diet mice and wild-type HFD mice, who displayed a sucrose preference (WT-Chow 65.43 ± 5.503 vs. WT-HFD 71.37 ± 3.797, p > 0.05; C/EBPβ Tg-Chow 62.62 ± 6.047 vs. C/EBPβ Tg-HFD 42.83 ± 3.961, p = 0.0441; Figure 4K), although there was no significant difference among different groups in pre-treatment of SPT (Figure 4J). All in all, these results confirmed that human C/EBPβ overexpression in mice neuron increased HFD-induced depression-like behaviors.

It has been reported that chronic stress leads to imbalance of the glutamatergic system, and dysregulation of glutamate signaling is increasingly considered to be a critical cause in mood disorders (Lee et al., 2012; Thompson et al., 2015; Duman et al., 2016). Here we hypothesized that C/EBPβ or HFD also caused glutamatergic system dysfunction. We treated WT or Thy1-C/EBPβ Tg male mice with HFD or chow diet for 12 weeks and performed electrophysiology, molecular experiments, staining and behavioral experiments (Figure 5A).

To investigate the role of C/EBPβ and HFD in glutamate neurotransmission, we first examined the input/output (I/O) curves in the CA1 region of hippocampal slices. The input/output curves were markedly reduced in the WT-HFD mice but with no significant change in the C/EBPβ+/− HFD mice (Figure 5B). In addition, C/EBPβ+/− HFD mice showed a normal provocation and maintenance of LTP in Schaffer collateral-CA1 compared to WT mice (Figure 5C; WT-Chow 1.55 ± 0.0478 vs. WT-HFD 1.126 ± 0.0785, p = 0.0085; WT-HFD 1.126 ± 0.0785 vs. C/EBPβ+/−-HFD 1.618 ± 0.1912, p = 0.0271; Figure 5D). To investigate whether C/EBPβ impairs glutamatergic neurotransmission via decreasing surface expression of glutamate receptors, we tested the expression of GluA1 and GluA2 by Western blotting. After exposure to HFD, C/EBPβ+/− HFD mice showed increased expression of surface and total GluA1 and GluA2 in the hippocampus versus WT-HFD mice, while the level of intracellular GluA1 displayed the opposite trend (Figures 5E–I; WT-Chow 1 ± 0.09442 vs. WT-HFD 0.5627 ± 0.0843, p = 0.0016; WT-HFD 0.5627 ± 0.0843 vs. C/EBPβ+/−-HFD 1.248 ± 0.0555 p < 0.0001; Figure 5F; WT-Chow 1 ± 0.1025 vs. WT-HFD 0.4871 ± 0.06233, p = 0.0001; WT-HFD 0.4871 ± 0.06233 vs. C/EBPβ+/−-HFD 1.139 ± 0.03759 p < 0.0001; Figure 5G).

In addition, immunostaining of hippocampus slices revealed that C/EBPβ+/− mice largely blocked the reduction of PSD95 compared to WT-HFD mice (Figure 5J; PSD95: WT-Chow 1.003 ± 0.0742 vs. WT-HFD 0.4185 ± 0.0601, p < 0.0001; WT-HFD 1.043 ± 0.0328 vs. C/EBPβ+/−-HFD 0.9396 ± 0.0609, p > 0.05; IL-6: WT-Chow 0.9944 ± 0.0469 vs. WT-HFD 2.747 ± 0.2132, p < 0.0001; WT-HFD 0.714 ± 0.1018 vs. C/EBPβ+/−-HFD 0.9728 ± 0.158, p > 0.05; Figure 5K), indicating that knockdown of CEBPB alleviated synaptic dysfunction. On the other hand, the expression of IL-6 decreased both in HFD and chow diet C/EBPβ+/− mice versus WT-HFD mice, suggesting that abolishing CEBPB reduced neuroinflammation (Figures 5J,K). Golgi staining displayed that deletion of C/EBPB significantly increased the dendritic spines and mushroom spines which are classically considered as mature dendritic spines in C/EBPβ+/− mice on both feeding conditions. By contrast, the reduced spines were only observed in WT-HFD mice (Figure 5L; WT-Chow 14.69 ± 0.835 vs. WT-HFD 7.769 ± 1.105, p < 0.0001; WT-HFD 15.54 ± 0.7127 vs. C/EBPβ+/−-HFD 13.85 ± 0.775, p > 0.05; Figure 5M; WT-Chow 5.154 ± 0.4507 vs. WT-HFD 2.385 ± 0.5609, p = 0.0002; WT-HFD 5.846 ± 0.191 vs. C/EBPβ+/−-HFD 5.154 ± 0.4213, p > 0.05; Figure 5N). Taken together, these results suggest that genetic ablation of CEBPB ameliorates HFD-induced synaptic plasticity impairment and neuro-inflammation in hippocampus.

To better understand the role of C/EBPβ in HFD-induced depression, we treated WT or Thy1-C/EBPβ Tg male mice with HFD or chow diet for 12 weeks. We found that HFD mice gained more weight than their chow-diet littermates regardless of WT or Thy1-C/EBPβ Tg male mice (Supplementary Figures S5A,B). ITT showed that both WT and Thy1-C/EBPβ Tg male mice had insulin intolerance after HFD treatment and GTT showed stronger glucose intolerance after HFD treatment than chow diet in both WT and Thy1-C/EBPβ Tg male mice (Supplementary Figures S5C,D). Then, we performed TST, FST and SPT. We found that HFD treatment made both WT and Thy1-C/EBPβ Tg male mice more depressive than chow diet mice, as indicated by the longer immobility time in TST and FST and lower sucrose preference in SPT (WT-Chow (12 W) 66.94 ± 5.669 vs. WT-HFD (12 W) 124.1 ± 15.71, p = 0.005; C/EBPβ Tg-Chow (12 W) 80.5 ± 8.382 vs. C/EBPβ Tg-HFD (12 W) 159.4 ± 12.82, p < 0.0001; Supplementary Figure S5E; WT-Chow (12 W) 69.06 ± 6.228 vs. WT-HFD (12 W) 139.5 ± 12.1, p < 0.0001; C/EBPβ Tg-Chow (12 W) 75.77 ± 5.797 vs. C/EBPβ Tg-HFD (12 W) 157.8 ± 12.72, p < 0.0001; Supplementary Figure S5F; WT-Chow (12 W) 63.68 ± 3.774 vs. WT-HFD (12 W) 42.85 ± 3.325, p = 0.0013; C/EBPβ Tg-Chow (12 W) 64.25 ± 4.505 vs. C/EBPβ Tg-HFD (12 W) 27.46 ± 2.939, p < 0.0001; Supplementary Figure S5H). These results indicated that 12 weeks of HFD treatment induced depression-like behaviors in both WT and Thy1-C/EBPβ Tg male mice.

To ascertain that neuroinflammation is the indeed initiating factor, we tested the influence of non-steroid anti-inflammatory drug (NSAID) Aspirin or Vehicle on HFD-induced depression-like behaviors in mice. WT and Thy1 C/EBPβ Tg male mice were fed with HFD for 12 weeks, and were treated Aspirin or vehicle for the last 4 weeks (Figure 6A). After the 3-month intervention, we collected mice tissue and quantified human CEBPB and mouse BDNF mRNA in mice hippocampi. We found no significant difference human CEBPB mRNA levels between Thy1-C/EBPβ Tg-HFD-Aspirin and Thy1-C/EBPβ Tg-HFD-Vehicle mice. However, Thy1-C/EBPβ Tg-HFD-Aspirin mice had higher hippocampal mouse BDNF mRNA level than that in Thy1-C/EBPβ Tg-HFD-Vehicle mice. These results indicate that inhibiting neuronal inflammation can affect the function of CEBPβ and thus up-regulated BDNF. (C/EBPβ Tg-HFD + Vehicle 1 ± 0.036 vs. C/EBPβ Tg-HFD + Aspirin 0.8825 ± 0.1328, p > 0.05; Figure 6B; WT-HFD + Vehicle 1 ± 0.1057 vs. WT-HFD + Aspirin 0.9497 ± 0.1817, p > 0.05; C/EBPβ Tg-HFD + Vehicle 0.5919 ± 0.1277 vs. C/EBPβ Tg-HFD + Aspirin 0.9871 ± 0.1332 p = 0.0358; Figure 6C).

Next, we measured pro-inflammatory cytokines in the hippocampus and serum by ELISA. In both tissues, Thy1-C/EBPβ Tg mice treated with aspirin had lower TNF-α, IL-1β, and IL-6 levels than their corresponding vehicle mice. (TNF-α: WT-HFD + Vehicle 1 ± 0.0999 vs. WT-HFD + Aspirin 0.7198 ± 0.1046, p = 0.0456; C/EBPβ Tg-HFD + Vehicle 1.514 ± 0.0933 vs. C/EBPβ Tg-HFD + Aspirin 0.942 ± 0.0674, p = 0.0009; IL-1β: WT-HFD + Vehicle 1 ± 0.1082 vs. WT-HFD + Aspirin 0.8884 ± 0.0820, p > 0.05; C/EBPβ Tg-HFD + Vehicle 2.566 ± 0.3666 vs. C/EBPβ Tg-HFD + Aspirin 1.184 ± 0.0641 p = 0.0002; IL-6: WT-HFD + Vehicle 1 ± 0.1407 vs. WT-HFD + Aspirin 0.7748 ± 0.0880, p > 0.05; C/EBPβ Tg-HFD + Vehicle 4.72 ± 0.3398 vs. C/EBPβ Tg-HFD + Aspirin 1.42 ± 0.1313 p < 0.0001; Figure 6D; TNF-α: WT-HFD + Vehicle 1 ± 0.1239 vs. WT-HFD + Aspirin 1.033 ± 0.1333, p > 0.05; C/EBPβ Tg-HFD + Vehicle 2.463 ± 0.0435 vs. C/EBPβ Tg-HFD + Aspirin 1.272 ± 0.0680, p < 0.0001; IL-1β: WT-HFD + Vehicle 1 ± 0.1716 vs. WT-HFD + Aspirin 0.792 ± 0.0901, p > 0.05; C/EBPβ Tg-HFD + Vehicle 1.754 ± 0.0555 vs. C/EBPβ Tg-HFD + Aspirin 1.363 ± 0.0824 p = 0. 027; IL-6: WT-HFD + Vehicle 1 ± 0.171 vs. WT-HFD + Aspirin 0.8113 ± 0.097, p > 0.05; C/EBPβ Tg-HFD + Vehicle 1.793 ± 0.0853 vs. C/EBPβ Tg-HFD + Aspirin 1.321 ± 0.1023 p = 0.0337; Figure 6E). In alignment with mRNA level, western blotting and ELISA experiments showed that Thy1-C/EBPβ Tg mice treated with aspirin had higher BDNF protein levels than those treated with the vehicle, while the two groups had similar human C/EBPβ levels (WT-HFD + Vehicle 1 ± 0.0640 vs. WT-HFD + Aspirin 1.139 ± 0.114, p > 0.05; C/EBPβ Tg-HFD + Vehicle 0.6168 ± 0.0322 vs. C/EBPβ Tg-HFD + Aspirin 0.9345 ± 0.0513 p = 0.0224; Figure 6F; C/EBPβ/Tubulin: C/EBPβ Tg-HFD + Vehicle 1 ± 0.0319 vs. C/EBPβ Tg-HFD + Aspirin 0.8684 ± 0.1052 p > 0.05; BDNF/Tubulin: WT-HFD + Vehicle 1 ± 0.1134 vs. WT-HFD + Aspirin 0.9874 ± 0.0547, p > 0.05; C/EBPβ Tg-HFD + Vehicle 0.4012 ± 0.0810 vs. C/EBPβ Tg-HFD + Aspirin 0.7935 ± 0.0699 p = 0.0187; Figure 6H).

To further confirm above molecular results, we then performed depression-like behavior tests and found that Thy1-C/EBPβ Tg mice treated with Aspirin showed fewer depressive behaviors. With less Namely, they had shorter immobility times immobility time in the TST and FST and a higher sucrose preference in SPT than those treated with the vehicle (WT-HFD + Vehicle 87.26 ± 3.970 vs. WT-HFD + Aspirin 66.91 ± 6.433, p > 0.05; C/EBPβ Tg-HFD + Vehicle 94.9 ± 10.62 vs. C/EBPβ Tg-HFD + Aspirin 143.9 ± 14.5 p = 0.008; Figure 6I; WT-HFD + Vehicle 88.54 ± 5.174 vs. WT-HFD + Aspirin 70.49 ± 6.431, p > 0.05; C/EBPβ Tg-HFD + Vehicle 82.77 ± 8.465 vs. C/EBPβ Tg-HFD + Aspirin 135.1 ± 15.08 p = 0.0038; Figure 6J; WT-HFD + Vehicle 46.01 ± 3.865 vs. WT-HFD + Aspirin 71.37 ± 3.747, p > 0.05; C/EBPβ Tg-HFD + Vehicle 62.62 ± 6.047 vs. C/EBPβ Tg-HFD + Aspirin 42.83 ± 3.961 p = 0.0441; Figure 6L). Finally, these results suggest that Aspirin alleviates HFD-induced depression-like behaviors in Thy1-C/EBPβ Tg mice.

To confirm that C/EBPβ promotes HFD-induced depression via downregulating its downstream factor BDNF, we directly injected AAV-BDNF or AAV-GFP into the hippocampus of WT or Thy1-C/EBPβ Tg male mice fed an HFD for 8 weeks and continued the diet for another 4 weeks (Figure 7A). The accuracy of the injection site and the expression of the target proteins was confirmed by immunostaining and Western blotting (Figures 7B–D; WT-HFD + AAV-GFP 1 ± 0.0802 vs. WT-HFD + AAV-BDNF 8.447 ± 0.7496, p < 0.0001; C/EBPβ Tg-HFD + AAV-GFP 0.9834 ± 0.07768 vs. C/EBPβ Tg-HFD + AAV-BDNF 8.25 ± 1.00, p < 0.0001; Figure 7E). Then, we performed the depressive behavior tests and found that Thy1-C/EBPβ Tg mice injected with AAV-BDNF had less immobility time in the TST and FST, and had a higher sucrose preference in the SPT than Thy1-C/EBPβ Tg mice injected with AAV-GFP (WT-HFD + AAV-GFP 88.34 ± 6.202 vs. WT-HFD + AAV-BDNF 63.49 ± 5.655, p < 0.0001; C/EBPβ Tg-HFD + AAV-GFP 138.4 ± 12.35 vs. C/EBPβ Tg-HFD + AAV-BDNF 80.71 ± 8.141, p = 0.0002; Figure 7F; WT-HFD + AAV-GFP 89.26 ± 5.063 vs. WT-HFD + AAV-BDNF 71.32 ± 5.302, p < 0.0001; C/EBPβ Tg-HFD + AAV-GFP 145.9 ± 11.54 vs. C/EBPβ Tg-HFD + AAV-BDNF 88.18 ± 6.542, p < 0.0001; Figure 7G; WT-HFD + AAV-GFP 48.55 ± 5.296 vs. WT-HFD + AAV-BDNF 59.88 ± 4.583, p < 0.0001; C/EBPβ Tg-HFD + AAV-GFP 25.22 ± 1.762 vs. C/EBPβ Tg-HFD + AAV-BDNF 61.91 ± 4.325, p < 0.0001; Figure 7I). These results confirmed that overexpression of BDNF in hippocampus alleviated HFD-induced depression-like behaviors in Thy1-C/EBPβ Tg mice.

The BDNF mimetic compound 7,8-DHF is a high-affinity TrkB agonist which can induce TrkB dimerization and autophosphorylation as well as activate downstream signaling (Jang et al., 2010; Chen et al., 2018). To confirm that BDNF and its major ligand-specific receptor TrkB mediates induces C/EBPβ-induced depression, we treated WT and Thy1-C/EBPβ Tg male mice with 7,8-DHF during the last 4 weeks of a 12-week HFD course a (Figure 8A). The Western blotting analysis showed that 7,8-DHF increased the phospho-TrkB (pTrkB)/TrkB ratio in wild-type and Thy1-C/EBPβ Tg male mice (Figure 8B; WT-HFD + Vehicle 1 ± 0.1539 vs. WT-HFD + 7,8-DHF 7.23 ± 1.736, p = 0.0325; C/EBPβ Tg-HFD + Vehicle 1.272 ± 0.07271 vs. C/EBPβ Tg-HFD + 7,8-DHF 8.309 ± 2.222, p = 0.0128; Figure 8C). In the behavioral tests, 7,8-DHF alleviated depression-like behaviors in Thy1-C/EBPβ Tg-HFD mice indicated as less immobility time in the TST and FST and higher sucrose preference in the SPT than their corresponding vehicle mice (WT-HFD + Vehicle 82.62 ± 7.868 vs. WT-HFD + 7,8-DHF 57.5 ± 4.455, p = 0.0002; C/EBPβ Tg-HFD + Vehicle 133.6 ± 10.3 vs. C/EBPβ Tg-HFD + 7,8-DHF 104.8 ± 6.785, p = 0.0186; Figure 8D; WT-HFD + Vehicle 87.62 ± 6.023 vs. WT-HFD + 7,8-DHF 71.15 ± 7.735, p < 0.0001; C/EBPβ Tg-HFD + Vehicle 168.1 ± 10.78 vs. C/EBPβ Tg-HFD + 7,8-DHF 131.5 ± 5.478, p = 0.0106; Figure 8E; WT-HFD + Vehicle 90.12 ± 5.831 vs. WT-HFD + 7,8-DHF 65.8 ± 4.171, p < 0.0001; C/EBPβ Tg-HFD + Vehicle 46.73 ± 5.703 vs. C/EBPβ Tg-HFD + 7,8-DHF 44.15 ± 4.915, p = 0.0081; Figure 8G).