SAG, a sonic hedgehog signaling agonist, alleviates anxiety behavior in high-fat diet-fed mice

Anxiety is a prevalent and disabling psychiatric disorder. Mitochondrial dysfunction due to the high-fat diet (HFD) was regarded as a risk factor in the pathogenesis of anxiety. The Sonic hedgehog (SHH) pathway was known to improve mitochondrial dysfunction through antioxidant and anti-apoptotic effects on some neuro- logical diseases. Nonetheless, its effect on anxiety has not been well studied. In this study, we aimed to explore whether SHH signaling pathway plays a protective role in anxiety by regulating mitochondrial homeostasis. SAG, a typical SHH signaling agonist, was administered intraperitoneally in HFD-fed mice. HFD-induced anxiety-like behavior in mice was confirmed using the open field and elevated plus maze tests. Immunofluorescence staining and Western blotting assays showed that the SHH signaling was downregulated in the prefrontal cortex neurons from HFD-fed mice. Electron microscopy results showed the mitochondria in the prefrontal cortex of HFD-fed mice were fragmented, which appeared small and spherical, and the area, perimeter and circularity of mitochondria were decreased. Mitofusin2 (Mfn2) and dynamin-related protein 1 (Drp1) were the key proteins involved in mitochondrial division and fusion. SAG treatment could rectify the imbalanced expression of Mfn2 and Drp1 in the prefrontal cortex of the HFD-fed mice, and alleviate the mitochondrial fragmentation. Furthermore, SAG decreased anxiety-like behavior in the HFD-fed mice. These findings suggested that SHH signal was neuroprotective in obesity and SAG relieved anxiety-like behavior through reducing mitochondrial fragmentation.


Introduction
Anxiety is the mind and body's reaction to stressful, dangerous or unfamiliar situations. Although a certain level of anxiety helps us stay alert and aware, people with anxiety disorders have intense, excessive and persistent worry and fear about everyday situations. Anxiety disorders with a prevalence of ~14% function in precursors of other mental disorders such as depression or even suicide, which is highly disabling and socioeconomically burdening (Charlson et al., 2019;Baxter et al., 2013;Craske et al., 2017).
Growing evidence accentuates the threat of obesity to central nervous system function and mental disorders. Clinical observations combined with animal models exhibits that the obesity increases the risk of anxiety disorders. The neurobiological interaction between obesity and anxiety are still not clearly understood, but studies indicate that mitochondrial dysfunction is involved in it (Filiou and Sandi, 2019;Pei and Wallace, 2018). It is well known that mitochondria play a pivotal role in the metabolism and maintenance of cellular homeostasis (Spinelli and Haigis, 2018;Picard et al., 2015). Any single neuron is equipped with a large number of mitochondria to meet its continuous energy demands. Meanwhile, mitochondria are critical regulators of neural structure and function. The structural and functional aberrations of the mitochondria have been evidenced in the central nervous system of the anxiety disorders and obesity (Andreazza and Nierenberg, 2018;Yang et al., 2021).
Mitochondria are dynamic organelles that undergo constant remodeling through the regulation of fission and fusion (Chan, 2006a). The imbalance between the fusion and fission rates leads to mitochondrial dysfunction (Bertholet et al., 2016). Excessive mitochondrial division or insufficient fusion results in mitochondrial fragmentation and reduces respiration and energy production, which is associated with the possibility of neuronal damage and cell death. The prefrontal cortex has been implicated in planning complex cognitive behavior, personality expression, decision making, and moderating social behavior. A strong association between the prefrontal cortex and emotion has been recognized. The effect of prefrontal cortex damage on emotional behavior has been confirmed with pertinent clinical and neuroimaging observations. Three clinical syndromes characterized by mood disorders, those are schizophrenia, major anxiety and depression disorders, which are dependent upon frontal system, especially prefrontal cortex dysfunction (Opel et al., 2021;Pandey et al., 2021;Wang et al., 2016). However, the relation and the mechanism of anxiety disorders to mitochondrial homeostasis in prefrontal cortex need to be further explored.
The sonic hedgehog (SHH) pathway mainly includes the ligand SHH, the receptor patched (PTCH), and the pathway activator smoothened (SMO) receptors. Briefly, the binding of SHH to PTCH results in the release and activation of SMO, and subsequently activation of GLI transcription factors. The SHH pathway is a well-characterized mitogen for progenitor cells in neural development (Fuccillo et al., 2006;Balaskas et al., 2012;Komada et al., 2008;Yabut and Pleasure, 2018). Furthermore, SHH signaling mediates neuroprotective and neurotrophic effects in many neurodegenerative diseases and in brain injury (Patel et al., 2017;Dellovade et al., 2006). Studies have shown that SHH pathway reduces mitochondrial damage and protects neurons against oxidative stress and apoptosis in autism, Parkinson's disease, stroke, and Down syndrome (Al-Ayadhi, 2012;Ghanizadeh, 2012;Tsuboi and Shults, 2002;Shao et al., 2017;Zhang et al., 2013;Jin et al., 2015;Ji et al., 2012;Huang et al., 2013;Das et al., 2013). Our previous research demonstrated that the SHH signaling pathway is downregulated in the hippocampus of mice fed a high-fat diet, which contributes to neuronal apoptosis (Qin et al., 2019). In mammals, Mitofusin2 (Mfn2) is essential for mitochondrial fusion, whereas dynamin-related protein 1 (Drp1) is responsible for mitochondrial fission. Studies shown SHH pathway protects endometrial cells against oxidative stress by suppressing Drp1 (Kaushal et al., 2018), but the activation of the SHH pathway increased mitochondrial abundance and activity in neurons (Yao et al., 2017). This implies that SHH signaling may regulate mitochondrial function by targeting mitochondrial dynamics.
In this study, we aim to further investigated whether SAG, a small molecule SHH pathway agonist, play a protective role in anxiety by regulating mitochondrial homeostasis. We displayed that SAG treatment rectified the imbalance of Mfn2 and Drp1 in the prefrontal cortex of the HFD-fed mice, and alleviated the structural changes of mitochondria.

Animal models and administration
Male C57BL/6 mice were obtained from the Model Animal Research Center of Nanjing University at 8 weeks of age. All mice were bred under specific pathogen-free conditions and acclimatized the laboratory conditions for a week.
Twenty mice were randomly divided into two groups: normal chow (NC, n = 10) group and high-fat diet group (HFD, n = 10). The high-fat diet contained 60 kcal% in fat, 20 kcal% in protein and 20 kcal% in carbohydrate (PD6001, SYSE BIO). The body weight of mouse was measured following a 6-h food withdrawal per 4 weeks (Qin et al., 2019;Kaushal et al., 2018). At the 20 weeks, all the mice were evaluated the anxiety-like behavior in open field and elevated plus maze. After behavioral testing, the SHH signal proteins in prefrontal cortex brain were detected using immunofluorescence staining and WB analyses.
Forty mice were randomly divided into four groups (10 mice per group): normal chow (NC), high-fat diet group (HFD), HFD + SHH signaling agonist group (HFD+SAG), and normal chow + SAG group (NC+SAG). The NC group and NC+SAG group mice were fed the normal diet, while the HFD group and the HFD + SAG group were fed the highfat diet. The SAG (>99% purity, Sigma, SML1314) was dissolved in sterilized saline water, and then administered between week 21 and week 32 via intraperitoneal injection at 10 mg/kg every 3 days, in the NC+SAG group and the HFD + SAG group mice. The dosage of SAG was selected according to previous behavioral and neurochemical studies (Tsuboi and Shults, 2002;Shao et al., 2017), and the administration interval was selected based on our preliminary experimental results (Qin et al., 2019;Su et al., 2020;Zhuang et al., 2019). Mice in the NC and HFD groups were injected with the same dose of sterilized normal saline. The body weight of mouse was measured every week. At the 32 weeks, the elevated plus maze and open field tests were used to measure anxiety-like behavior. After behavioral testing, the glucose tolerance tests were performed and the brain samples were collected for WB, IHC, electron microscopy and histology.
All experimental protocols were in compliance with the general guidelines of the National Institutes of Health for the Care and Use of Laboratory Animals in Scientific Investigations and were approved by the Laboratory Animal Ethics Committee of Xuzhou Medical University.

Elevated plus maze test
The elevated plus-maze is a test of anxiety-like behavior that uses the natural reluctance of rodents to explore open spaces (Iñiguez et al., 2014;Han et al., 2020;Samad et al., 2018). The elevated plus maze consisted of four arms (two open without walls and two enclosed by 15.25-cm high walls) 30 cm long and 5 cm wide. Each mouse was gently placed at the junction of the open and closed arms, with its head oriented toward the open arm (10 mice per group). They were allowed to explore the maze freely for five minutes. The excrement was cleaned, and the apparatus was scrubbed with 70% ethanol after each test to eliminate any residual olfactory cues from the previously tested mouse.

Open field test
Open field test is usually used to analyze exploratory, stress, and anxiety behavior but also to measure locomotor activity levels into the arena. The test was performed as previously reported (Tian et al., 2018;Yang et al., 2020). A square box made of dark opaque Plexiglas (40 cm × 40 cm × 40 cm) with the floor divided into 16 equal segments was used. Each mouse was gently placed in the middle of the field, and allowed to acclimate for 3 min, and then to freely explore the field for 5 min (10 mice per group). Excrements were cleaned, and the apparatus was scrubbed with 70% ethanol between each test to eliminate any residual olfactory cues from the previously tested mouse. The total traveled distance, zone transition number, distance in central zones (%), and the time in central zones (%) were recorded. The parameters were acquired and processed using the Smart 3.0 system (Panlab, Spain).

Body weight and glucose tolerance analysis
The body weight of mouse (10 mice per group) was measured weekly following a 6-h food withdrawal. Glucose tolerance test (10 mice per group) was performed by intraperitoneal glucose injection (1.5 g/kg) after 6 h fasting. The blood glucose was measured immediately before and 15, 30, 60, 90 and 120 min after glucose administration. The blood was collected from the tail vein, and the blood glucose levels were measured with glucose meter (7600 P, Bayer).

Collection of brain slices
The mice were deeply anesthetized and transcardially perfused with normal saline, followed by 4% paraformaldehyde (3 mice per group). The brains were removed, post-fixed in 4% paraformaldehyde for 4 h, then incubated in 20% sucrose overnight at 4 • C, followed by incubation in 30% sucrose overnight at 4 • C, prior to being embedded in tissue freezing medium (OCT compound, Leica). A series of 12-μm thick, parallel frozen coronal sections were cut and mounted on gelatinized slides. The sections were then stored at − 70 • C for immunohistochemistry staining.

Transmission electron microscopy
After anesthesia, the mice were perfused with normal saline and perfused with 4% paraformaldehyde (3 mice per group). About 1 mm 3 of the prefrontal cortex tissue was fixed in 2.5% glutaraldehyde for 4 h, rinsed in PBS for 4 h, post-fixed in 1% osmic acid for 2 h, and rinsed in distilled water for 3 h. The dehydration procedure was as follows: 50% ethanol-70% ethanol-80% ethanol-90% ethanol-90% ethanol + 90% acetone-90% acetone-100% acetone, 15 min for each step. The was then immersed using anhydrous acetone + epoxy resin (2:1) for 3.5 h followed by anhydrous acetone + epoxy resin (1:2) for 3.5 h. Sample embedding and polymerization was performed using epoxy resin per the following parameters: 37 • C for 24 h, 45 • C for 24 h, and 60 • C for 24 h. Ultra-thin sections (70 nm) were cut with an ultramicrotome and then stained with uranium acetate for 30 min and lead citrate for 10 min. The sections were observed (×4400) and imaged using a transmission electron microscope (FEI, Tecnai G2 Spirit Twin). The mitochondria were detected and analyzed in six microscopic fields per sample. The area, perimeter, and circularity of individual mitochondria were further analyzed using Image J software.

Tissue homogenization
The prefrontal cortex includes a small area within and dorsal to the rhinal sulcus and a comparatively larger region within the medial half of the anterior cortex (Premachandran et al., 2020;Le Merre et al., 2021). The mouse (3 mice per group) prefrontal cortex was homogenized in ice-cold RIPA lysis buffer (Beyotime, P0013B) supplemented with complete EDTA-free protease inhibitor cocktail and PhosSTOP phosphatase inhibitor(Beyotime，ST506). The homogenate was sonicated six times for 4 s at 6 s intervals on ice and then centrifuged at 12,000 g for 20 min at 4 • C. The supernatant was collected, and the protein concentration was quantitated using the BCA assay (Beyotime, P0009). The samples were prepared for Western blotting and biochemical analyses.

Evaluation of CAT, SOD activities and MDA concentration
The activities of antioxidant enzymes such as catalase (CAT) and superoxide dismutase (SOD) as well as the concentration of malondialdehyde (MDA), as the indicator of oxidative stress, were measured in the homogenates of prefrontal cortex (Liu et al., 2012;Zhu et al., 2014). The test was performed according to the manufacturer's instructions, and detected with the spectrophotometer (BioTek, Synergy2). The MDA content [nanomoles per milligram protein (nmol/mg protein)] was determined using a thiobarbituric acid assay kit (Nanjing Jiancheng Bioengineering Institute, A003-1). Optical density was measured at 532 nm. SOD activity was determined using a xanthine oxidase assay kits (Nanjing Jiancheng Bioengineering Institute, A001-3) and expressed as units per milligram of protein (U/mg protein). The optical density was measured at 450 nm wavelength. CAT activity was measured using a colorimetric assay kit (Nanjing Jiancheng Bioengineering Institute, A007-1) in units per milligram of protein (U/mg protein). The optical density was measured at 405 nm wavelength.

Statistical analysis
Statistical analyses were performed using GraphPad Prism version 6. The behavioral data are shown as median, quartiles and extremums. The other quantitative data are expressed as mean ± standard deviation. The Kolmogorov-Smirnov test was used to test if the distribution of the data was Normal Distribution. The body weight and blood glucose data were analyzed by two-way repeated measure ANOVA. The other data in the manuscript were analyzed by Student's unpaired t-test and two-way ANOVA. Differences were considered significant at P < 0.05.

HFD-fed enhanced anxiety-like behaviors in mice
A fat-enriched diet is the major cause of the prevalence of metabolic disorders in the population. Our data showed an increase in mice weight after 4 weeks of HFD feeding, and the HFD-fed mice were significantly heavier than the negative control (NC) mice by 20 weeks of HFD feeding (Fig. 1A). Growing evidence showed the obesity affected the function of central nervous system and mental disorders. After 20 weeks of HFD, the elevated plus maze and open field test were performed to verify whether obesity led to anxiety-like behavior in mice. In the elevated plus maze test, the latency to enter the open arm was longer, and the number of entering the open arms was significantly lower in the HFD group than that of NC group. Meanwhile, the HFD-fed mice traveled less distance (%) and kept less time in the open arms (%) (Fig. 1B and D). In the open field task, HFD feeding group was resulted in a significant reduction in the total traveled distance, percentage of distance (%), percentage of time (%) in the central area, and the zone transition number (Fig. 1C and E).

SHH and SMO were downregulated in the prefrontal cortex of HFDfed mice
The SHH pathway mediates neuroprotective and neurotrophic effects in a variety of neurological diseases. We first investigated the effect of HFD on the activation of SHH pathway in mouse cortex. Western blotting analysis ( Fig. 2A and B) revealed that the expression of SHH and SMO, as the key proteins of the SHH pathway, was significantly lower at 20 weeks in the prefrontal cortex of HFD-fed mice compared with the NC mice. The results from immunofluorescence staining further demonstrated SMO was co-expressed with neuronal marker NeuN (Fig. 2C).

The SAG improved the glucose metabolism indexes in HFD-fed mouse
The fat-enriched diet is the major cause of the marked prevalence of metabolic disorder in the population. Our data showed the HFD-fed mice were significantly heavier than that of the NC group mice at the 32 weeks (Fig. 3 A). In the glucose tolerance test, the mice received the HFD for 32 weeks showed significantly higher glucose levels than the NC group (Fig. 3B). However, intraperitoneal injection of SAG to HFD mice caused significant reduction in the body weight gain and improved the glucose tolerance.

The SAG upregulated the SHH in the prefrontal cortex of HFD-fed mice
SAG, the agonist of SHH pathway, binds to SMO, then upregulates and activates the SHH pathway. Western blotting analysis showed that SHH expression was downregulated in the prefrontal cortex of HFD-fed mice compared with that of the NC group, at the 32 weeks, SAG treatment increased the expression of SHH in the prefrontal cortex of HFDfed mice ( Fig. 4A and B). The results of immunofluorescence staining further demonstrated SHH was mainly co-expressed with neuronal marker NeuN (Fig. 4C). These showed the SAG treatment upregulated SHH in the neuron of the HFD-fed mice.

SAG alleviated anxiety-like behaviors in HFD-fed mice
A number of studies show that SAG has a protective role against neurodegenerative diseases and brain injury. The SAG was administered from 21 weeks to 32 weeks of NC and HFD feeding. To measure the effect of SAG on anxiety-like behaviors in mice, we employed the elevated plus maze test and the open field task at 32 weeks. In the elevated plus maze test, compared with the HFD group mice, HFD+SAG group mice took less time to enter the open arm, traveled more distance (%), kept more time in the open arms (%) and the number of entering the open arms was also higher (Fig. 5A, B, C, D and E). Then, the open field task was further performed. The HFD+SAG group mice displayed significant advantage in total distance traveled, the percentage of distance (%), and percentage of time (%) in the central area (Fig. 6A, B and C), compared with the HFD group mice. The zone transition number of   (A) Two-way repeated ANOVA of body weight with a subject factor of treatment (HFD or SAG) and a within subject factor of week was separately applied to the biologically meaningful 4 times of data between NC and HFD, between NC and HFD + SAG, between HFD and HFD + SAG, between NC and NC + SAG, These revealed significant interactions between NC and HFD (F (8, 144) = 35.9, P＜0.001), between NC and HFD + SAG (F (8, 144) = 16.761, P＜0.001), between HFD and HFD + SAG (F (8, 144) = 7.941, P＜0.001), with no main effect of SAG in the comparison between NC and NC + SAG (F (4.597, 82.748) = 0.376, P = 0.849). Each two-way repeated ANOVA was followed by post-hoc test of Holm's multiple comparisons, * P＜0.05, * * P＜0.01, * ** P＜0.001, compared with NC and ★ P＜0.05, ★★ P＜0.01, ★★★ P＜0.001, compared with HFD + SAG. (B) The blood glucose concentration in the glucose tolerance tests at the 32 weeks. Two-way repeated ANOVA of blood glucose with a subject factor of treatment (HFD or SAG) and a within subject factor of time (min) was separately applied to the same 4 times of data as shown above. These revealed significant interactions between NC and HFD (F (2.722, 49) = 6.114, P = 0.002), between HFD and HFD + SAG (F (5, 90) = 6.02, P＜0.001), but not between NC and HFD + SAG (F (5, 90) = 0.558, P = 0.732), between NC and NC + SAG (F (3.303, 59.449) = 1.290, P = 0.286). Accordingly, post-hoc test was performed by Holm's multiple comparisons; * ** P＜0.001, compared with the NC group and ★★★ P＜0.001, compared with the HFD+SAG group.
HFD+SAG group mice was also more than that of HFD group mice ( Fig. 6D and E). These data showed SAG treatment alleviated anxietylike behavior of HFD-fed mice in the elevated plus maze and the open field task.

SAG inhibited the neuronal apoptosis in the prefrontal cortex of HFDfed mice
We firstly investigated whether SAG could reduce the apoptosis of neurons in the prefrontal cortex of HFD-fed mice. We detected the changes of proteins related to apoptosis, such as Bcl-2, Bak and the Cleaved-caspase3. The results showed that the anti-apoptotic protein Bcl-2 was downregulated significantly while pro-apoptotic protein Bak were upregulated obviously in the prefrontal cortex of HFD-fed mice.
Meanwhile, the Cleaved-caspase3 were significantly upregulated in HFD-fed mice, compared with the NC group mice. After treatment with SAG, the expression of Bcl2 was increased, and the expression of Bak and Cleaved-caspase3 was decreased in the prefrontal cortex of HFD+SAG mice, compared with the HFD group ( Fig. 7A and B). Together, these data imply that SAG has neuroprotective effects against HFD-induced neuronal damage.

SAG alleviated mitochondria damage in the prefrontal cortex of HFD-fed mice
Changes in mitochondrial morphology have been linked to the apoptosis of neurons. In response to apoptotic stimulus, it has been demonstrated that cells undergo mitochondrial excessive division Relative density analysis of the protein expression. Two-way AVOVA with the subject factors of diet and drug (SAG) was performed for the data of SHH immunoreactivity. The main effect of diet was F (1, 56) = 11.901, P = 0.001 and that of drug was F (1, 56) = 1.300, P = 0.259, but with a significant interaction between diet and drug (F (1, 56) = 11.035, P = 0.002). Post-hoc test was performed by Holm's multiple compensation; * ** P＜0.001, compared with the NC group, and ★★★ P＜0.001, compared with the HFD+SAG group. (C) The immunofluorescence of SHH and NeuN immunoreactivities were photographed in the cerebral cortex. Scale bar: 25 µm.
generating fragmented mitochondria (Brooks et al., 2011). The results from electron microscopy showed that the mitochondria appeared small and spherical in the prefrontal cortex of HFD mice. The mitochondria in SAG+HFD group were rod-shaped, and the cristae of mitochondria were arranged in an orderly manner (Fig. 8A). Statistical analysis demonstrated that the area, perimeter, and circularity of mitochondria in HFD group were lower than those of NC groups, these implied that HFD-fed led to mitochondrial fragmentation in the prefrontal cortex of mice. After treatment with SAG, the area, perimeter, and circularity of mitochondria in SAG+HFD group mice were increased compared with the HFD group mice (Fig. 8B, C and D). This showed the SAG alleviated the mitochondrial fragmentation induced by HFD-fed.

SAG regulated Mfn2 and Drp1 in the prefrontal cortex of HFD-fed mouse
An imbalance in the cycles of fission and fusion can disturb the overall morphology causing mitochondria to appear as small and fragmented. Mfn2 is the key factor for mitochondrial fusion, whereas Drp1 is indispensable for mitochondrial fission. Western blotting analysis showed that Drp1 expression was significantly upregulated in the prefrontal cortex of HFD mice compared with that in the cortex of the NC group, whereas Mfn2 expression was significantly lower. This imbalance between the Drp1 and Mfn2 in HFD mice may lead to mitochondrial fragmentation. Interestingly, the SAG treatment reversed the expression of Mfn2 and Drp1 in the prefrontal cortex of HFD-fed mice ( Fig. 9A and   Fig. 6. SAG ameliorated the anxietylike behavior of the HFD-fed mice in the open-field test. (A) Two-way AVOVA with the subject factors of diet and drug (SAG) was performed for the data of the total travelled distance. The main effect of diet was F (1, 36) = 22.5, P < 0.001 and that of drug was F (1, 36) = 13.1, P < 0.001 with a significant interaction between diet and drug (F (1, 36) = 14.2, P < 0.001). (B) The data of the zone transition number were analyzed similarly. The main effect of diet was F (1, 36) = 33.54, P < 0.001 and that of drug was F (1, 36) = 9.34, P = 0.004 with a significant interaction between diet and drug (F (1, 36) = 15.01, P < 0.001). (C) The data of the time in central area were analyzed as described above. The main effect of diet was F (1, 36) = 37.7, P < 0.001 and that of drug was F (1, 36) = 25.9, P < 0.001 with a significant interaction between diet and drug (F (1, 36) = 26.7, P < 0.001). (D) The data of the distance in central area were similarly analyzed. The main effect of diet was F (1, 36) = 13.46, P < 0.001 and that of drug was F (1, 36) = 6.54, P = 0.015 with significant interaction between diet and drug (F (1, 36) = 7.00, P = 0.012). (A, B, C, D) Post-hoc test was followed in all the above analyses by Holm's multiple comparisons; * ** P ＜0.001, compared with the NC group and

SAG regulated MDA, CAT and SOD in the prefrontal cortex of HFDfed mice
Because of impaired mitochondrial function aggravating oxidative stress, the indicator of oxidative stress such as antioxidant enzymes catalase (CAT), superoxide dismutase (SOD) as well as malondialdehyde (MDA)were detected in the prefrontal cortex of mice. It is well known that SOD and MDA are dynamically balanced under physiological conditions, which controls the content of reactive oxygen species. The results showed that HFD-fed upregulated the MDA production (Fig. 10A) and decreased the activities of SOD and CAT in the prefrontal cortex of mice ( Fig. 10B and C). The activity of CAT, SOD and MDA were reversed in the SAG+HFD group, compared with the HFD group. These implied the SAG alleviated the oxidative stress in the prefrontal cortex of HFDfed mice.

Discussion
Anxiety is a state of uneasiness and enhanced vigilance in the absence of an immediate threat, frequently accompanied by behavioral and cognitive distress responses. Anxiety constitutes the most common mental disorders with a 12-month prevalence of ~14%, posing a major human and economic burden in modern society (Charlson et al., 2019).
Obesity is now a worldwide public health problem and extensively impacts on physiology and health. Growing evidence is accentuating the threat of obesity to central nervous system function and risk of psychiatric illness. A positive relationship is observed with anxiety: obesity increases the odds of an anxiety disorder or anxiety symptoms by 30%  and 40% (Fulton et al., 2022;Milaneschi et al., 2019). However, the mechanism remains unclear. Clinical observations combined with rodent models of obesity exhibiting anxiety-like behaviors proves valuable for uncovering the neural mechanisms. Unhealthy diets that include excess saturated fat and sugar intake promote obesity, metabolic dysfunction, neuroinflammation, and mental health impairments (Kim et al., 2019;Hollis et al., 2018). Our data consistently revealed that mice treated with HFD for 32 weeks were significantly heavier than the NC mice. Based on the results of the open field and elevated plus maze tests, HFD exactly induced anxiety-like behavior in mice. The mechanism of neurobiological interaction between obesity and anxiety should be further investigated.
Mitochondria plays a key role in energy production and in the maintenance of cellular stability by maintaining reactive oxygen species levels, modulating Ca 2+ levels, and regulating apoptosis. Studies have sparked the appreciation for the role of mitochondria in many intracellular processes, synaptic plasticity, and cellular resilience (Todorova and Blokland, 2017). There is also increasing evidence indicating that mitochondrial dysfunction is a hallmark of neurological diseases including anxiety, depression, Alzheimer's disease, Parkinson's disease and Huntington's disease (Bansal and Kuhad, 2016;Lin and Beal, 2006;Wang et al., 2020;Sironi et al., 2020;Bose and Beal, 2016;Shirendeb et al., 2011;Chistiakov et al., 2014;Ben-Shachar and Ene, 2018), suggesting mitochondria-targeted treatments become a plausible treatment avenue for these diseases. Accumulating data confirm the contribution of brain mitochondria to stress-related pathologies and psychiatric disorders (Picard et al., 2018;Hollis et al., 2015). A bidirectional interplay between anxiety and brain mitochondria has been revealed. Substantial observations indicate broad range of alterations in mitochondria and brain energy metabolism as a critical feature of anxiety. Conversely, anxiety symptoms in humans suffering from mitochondrial disorders. Our results also manifested the mitochondria in the prefrontal cortex of HFD-fed mice were fragmented with small and spherical shape, decreased the area, perimeter, and circularity of mitochondria. The levels of oxidative stress and apoptosis were increased in the prefrontal cortex of HFD-fed mice that were correlated with the mitochondrial fragmentation. These imply that the obesity is coupled to various structural and functional changes in mitochondria of the brain that are remarkably similar to those observed in anxiety disorders.
Mitochondria support cellular adaptation to a variety of challenges such as stress, through their capacities to fusion-fission dynamics (Chan, 2006b). This dynamic maintains the population of mitochondria, assists mitochondria interact with other organelles (e.g., endoplasmic reticulum), and promotes mitochondria migrate to diverse cell locations. The mitochondrial dynamic is increasingly revealed to be crucial for a broad range of neuronal processes (e.g., neuronal growth and sprouting, synaptic transmission, neuronal plasticity, and connectivity). Cells with a high fusion-to-fission ratio have fewer mitochondria, and these mitochondria are long and highly interconnected. Conversely, cells with a low fusion-to-fission ratio have numerous mitochondria that appear as small spheres or short rods, often referred to as 'fragmented mitochondria' (Westermann, 2012). The imbalance between fusion and fission is important factor associated with mitochondrial dysfunction. Neurons are metabolically active cells with high energy demands, accordingly, is highly dependent on mitochondria. As a result, neurons are particularly vulnerable to injury and death from mitochondrial dysfunction. Our results from electron microscopy revealed that the mitochondria in neurons from prefrontal cortex of HFD-fed mice was visible vacuolated, fragmented that appeared small and spherical，which implied the HFD treatment led to a low fusion-to-fission ratio of the mitochondria. Mitofusins (Mfn1 and Mfn2) are required for mitochondrial membrane fusion, whereas the dynamin-related protein Drp1 is the key regulator of mitochondrial fission. Our results showed that HFD induced an imbalance between Drp1 and Mfn2. The Drp1 expression was upregulated and the Mfn2 expression was significantly lower in the prefrontal cortex of HFD-fed mice. The results of electron microscopy further displayed the size of mitochondria in the HFD group was small, the cristae of mitochondria were decreased or in irregularly arranged. Furthermore, the density of the matrix in the mitochondria was also lower.
The SHH signaling pathway, which includes SHH, PTCH, SMO and GLI, is considered essential in the developing nervous system (Robbins et al., 2012). SHH signaling promotes the proliferation of neural stem cells, dendritic spine density, and synaptic plasticity, and has neurotrophic effects in many neurological injuries. In animal models of Parkinson's disease and Down syndrome, SHH protects rat cortical neurons from oxidative stress and apoptosis (Tsuboi and Shults, 2002;Das et al., 2013). Evidence showed that stimulating the SHH pathway promotes repair and reduces infarct size after stroke in mice (Zhang et al., 2013). Here, we observed that SHH signaling was downregulated in the prefrontal cortex of HFD-fed mice. The agonist of SHH (SAG) was administrated to ameliorate the effect of the SHH pathway on the anxiety-like behavior of HFD-fed mice. It is known that SAG is a derivative of chloro-benzo-thiophene and can cross the gut, placenta, and blood-brain barrier. SAG binds to and activates SMO, thus upregulating the SHH pathway and stimulating the activities of SHH downstream in vitro and in vivo (Rimkus et al., 2016;Hadden, 2014). SAG have a protective role were measured in the prefrontal cortex and compared by two-way AVOVA with the subject factors of diet and drug. Significant interactions between diet and drug were detected for all the enzyme activities; F (1, 44) = 187, P < 0.001 for SOD，F (1, 20) = 35.2, P < 0.001 for CAT, and F (1, 44) = 24.6, P < 0.001 for MDA. Thus, post-hoc test was performed to detect differences between individual pairs by Holm's multiple comparisons; * ** P＜0.001, compared with the NC group and ★★★ P＜ 0.001, compared with the HFD+SAG group.
against neurodegenerative diseases and brain injury (Das et al., 2013;Nguyen et al., 2018). Studies have demonstrated the beneficial effects of SAG against neonatal cerebellar injury and it improves Down syndrome-related brain structural deficits in mice. Furthermore, our previous research also demonstrated SAG treatment reduced the apoptosis of hippocampal neurons and ameliorated the cognitive impairment in HFD-fed mice. Here, we also confirmed that SAG relieved anxiety-like behavior in HFD-fed mice. Studies have shown that SAG increases mitochondrial abundance and activity in neurons (Yao et al., 2017). We observed that SAG alleviated mitochondrial structural changes and reversed the expression of Mfn2 and Drp1 in the prefrontal cortex of HFD-fed mice. Furthermore, mitochondria are the chief producer of reactive oxygen species and play a key role in the regulation of cell apoptosis. Our results displayed SAG downregulated MDA production, increased SOD and CAT activities, and decreased the expression of Cleaved caspase3 and Bak in the prefrontal cortex of HFD-fed mice. The result also showed the SAG halt weight gain and improve the glucose tolerance. The SHH signaling regulates the functions during pancreas development and in adult tissue (Kayed et al., 2006). The SHH signaling proteins are expressed in the β cells of islets. The inhibitor of the signaling, cyclopamine, reduces insulin production in a dose-dependent manner. It also suggests that SAG may play an indirect neuroprotective role by improving metabolism.
In conclusion, our study showed that HFD downregulates the SHH signaling pathway and contributes to the mitochondrial dysfunction in the prefrontal cortex of mice, which are result in anxiety-like behavior in obesity mice. SAG relieved anxiety-like behavior through reducing mitochondrial fragmentation by rectifying the imbalance between Drp1 and Mfn2 in the prefrontal cortex of HFD-fed mice. Thus, our findings emphasize that SAG play roles in anxiety-like behavior and neuroprotection in obesity.

Funding
This work was supported by the National Natural Science Foundation of China (81971179 to Xiaomei Liu), the Natural Science Foundation of Jiangsu Province (BK20191463 to Xiaomei Liu), Jiangsu Commission of Health (Z2019035 to Feng Zhou), Jiangsu Provincial Department of Education (20KJA320004 to Feng Zhou), the Priority Academic Program Development of Jiangsu Higher Education Institutions (PAPD) and the Research Foundation of Jiangsu Province Key Laboratory of Immunity and Metabolism (JSKIM201704 to Dexu Sun).

CRediT authorship contribution statement
Xiaomei Liu and Feng Zhou designed the experiments and revised the manuscript; Dexu Sun, Suping Qin, Jiaxin Deng, Yifan Wang, Jinyu Xie and other authors performed the experiment; Dexu Sun analyzed the data and wrote the manuscript.