NSC 178886

Paeoniflorin exerts antidepressant-like effects through enhancing neuronal FGF-2 by microglial inactivation

Jie Cheng a, Min Chen a, Hui-Qi Wan b, Xue-Qin Chen c, Cheng-Fu Li c, Ji-Xiao Zhu d, Qing Liu a, Guang-Hui Xu b,**, Li-Tao Yi a, e,*

A B S T R A C T

Ethnopharmacological relevance Paeonia lactiflora is a famous Traditional Chinese medicine widely used for immunological regulation. Paeoniflorin, the main component of Paeonia lactiflora, exerts neuroprotective and antidepressant-like effects in rodents.
Aim of the study: Fibroblast growth factor 2 (FGF-2) is essentially required in the central nervous system as it acts as both a neurotrophic factor and an anti-inflammatory factor participating in the regulation of proliferation, differentiation and apoptosis of neurons in the brain. However, it is unclear whether paeoniflorin could exert antidepressant effects via regulating FGF-2.
Materials and methods: In the present study, the effects of paeoniflorin were evaluated in depressive mice induced by the endotoXin lipopolysaccharide (LPS) injection.
Results: The results showed that paeoniflorin markedly increased sucrose preference and reduced immobility time in LPS mice, indicating antidepressant effects. Consistent with the results from molecular docking showing paeoniflorin antagonizes TLR4, NF-κB and NLRP3, the biochemical analysis also indicated paeoniflorin inhibited TLR4/NF-κB/NLRP3 signaling, decreased proinflammatory cytokine levels and microglial activation in the hippocampus of LPS induced mice. In addition, the levels of neuronal FGF-2 and the density of dendritic spine were improved by paeoniflorin. More importantly, the FGFR1 inhibitor SU5402 prevented the antidepressant effects of paeoniflorin and blocked the neuroinflammatory and neurogenic regulatory effects of paeoniflorin, indicating that FGF-2/FGFR1 activation was required for the effects of paeoniflorin.
Conclusion: Taken together, the results demonstrate that paeoniflorin exhibits neuroprotective and antidepressant effects in mice, which may be mediated by activating neuronal FGF-2/FGFR1 signaling via the inhibition of microglial activation in the hippocampus.

Keywords: Paeoniflorin FGF-2 Microglia Neuroinflammation Antidepressant

1. Introduction

Depression is a mental illness characterized by emotional, cognitive, psychomotor, and neurasthenic symptoms that spread throughout all aspects of life, undermining individual family and personal relation- ships, work adjustments, and general health (Catena-Dell’Osso et al., 2011). The factors that contribute to depression are very complex, which causes difficulty in understanding the specific pathogenesis of depression. To date, it has been widely accepted that the typical causes of depression are related to genetic, immunological, neuroendocrine, neurochemical, and social and environmental factors (Jesulola et al., 2018).
The causal relationship between these diseases and inflammation is under extensive investigation (Casaril et al., 2019). Neuroinflammation is typically involved in the pathogenesis of neurodegenerative diseases, such as Alzheimer’s disease and Parkinson’s disease (Phan and Malkani, 2019). After acute brain injury or infection, there is a complex inflam- matory response that involves activation of microglia and astrocytes, and increased production of cytokines, chemokines, acute phase pro- teins, and complement factors, antibodies and T lymphocytes may also be involved in the response (Buckwalter and Wyss-Coray, 2004). Microglial activation is necessary and critical for host defense, but excessive activation of microglia is harmful if there are acute brain damage and infection (Polazzi and Contestabile, 2002). The inflamma- tory response in neurodegenerative diseases is continuous and pro- longed and may have devastating effects (Wyss-Coray and Mucke, 2002). In patients with depression, cerebrospinal fluid or serum con- centrations of proinflammatory cytokines such as interleukin-1β (IL-1β), interleukin-6 (IL-6) and tumor necrosis factor-α (TNF-α) increase (Enache et al., 2019; Tajfard et al., 2014), indicating the relationship between depression and inflammation.
Fibroblast growth factor 2 (FGF-2) is a multifunctional growth factor that plays a crucial role in cell development. FGF-2 has been previously implicated in several biological processes during diseases, such as mental diseases (Katsouri et al., 2015). More importantly, several studies found that exogenous FGF-2 infusion could exert an antide- pressant effect (Elsayed et al., 2012; Wang, et al., 2018a), which was partly due to the regulation of FGF-2 in neuronal proliferation and dif- ferentiation. Similarly, FGF-2 was required for the therapeutic effects of the antidepressant fluoXetine (Simard et al., 2018). Our recent study showed that the neuroinflammation induced by LPS decreased FGF-2 levels (Chen et al., 2020). On the other hand, FGF-2 was shown to inhibit microglial activation and proinflammatory cytokine release (Tang et al., 2018). In this way, it is rational to suggest that the abnor- mality of neuroinflammation/FGF-2 is involved in the pathophysiology of depression.
Paeonia lactiflora is one of the famous Traditional Chinese Medicine which is widely used for immunological regulation. Several traditional formulas, such as Chaihu Shugan San and Xiaoyao San, which contain Paeonia lactiflora, have been used for depression treatment for a long time (Chen et al., 2018; Liu, et al., 2019b). Paeoniflorin, the main and active component of Paeonia lactiflora, has been widely studied as an anti-inflammatory and neuroprotective agent (Qiu, et al., 2013a). More importantly, paeoniflorin has been verified as an antidepressant in ro- dents (Qiu, et al., 2013a,b). The antidepressant activity of paeoniflorin may be due to the overexpression of brain-derived neurotrophic factor (BDNF), recovery of neural stem cell proliferation, activation of ERK/- CREB signaling and inhibition of inflammatory cytokines in rodents (Chen et al., 2019; Hu et al., 2019; Li et al., 2017; Liu, et al., 2019a; Zhong et al., 2018). Although there was evidence showing that micro- glial inactivation mediated the antidepressant effects of paeoniflorin (Li et al., 2017; Tian et al., 2021), microglial/FGF-2 interaction in the an- tidepressant effects of paeoniflorin are unclear. Therefore, in the present study, the anti-inflammatory targets were firstly simulated by molecular docking. Then, the mechanism of paeoniflorin involving microglia/FGF-2 regulation was investigated in the endotoXin lipo- polysaccharide (LPS) induced depression. Furthermore, to elucidate whether FGF-2 enhancement was required for the effects of paeoniflorin, the FGFR1 antagonist SU5402 was pretreated prior to paeoni- florin administration.

2. Materials and methods

2.1. Animals

The ICR mice (male, 20 2 g, 8 weeks old) were provided by Shanghai Slac Animal Center, PR China. During the experiment, four mice were placed in one cage (320 180 160 mm) under a normal 12h/12 h circadian rhythm. The light was turned on at 08:00 a.m. and turned off at 08:00 p.m. The room temperature was set at 22 2 ◦C and the relative humidity was set at 55 5%. There was a one week adaptation period for animals before the start of the current experiment. During the entire experiment, animals were given food and water unless otherwise stated. The entire experiment was approved by the Animal Committee of Huaqiao University (A2019017) which complied with the guidelines of the China Council on Animal Care.

2.2. Reagents

Paeoniflorin (purity>98% verified by HPLC, P101691) was pur- chased by Aladdin (Shanghai, China). IL-1β, IL-6 and TNF-α ELISA kits were purchased from Boster (Wuhan, China). The FD Rapid GolgiStain™ Kit was purchased from FD NeuroTechnologies (Columbia, USA). FluoXetine, LPS, DAPI and primary β-actin antibodies were purchased from Sigma (St. Louis, USA). The antibodies for FGF-2 and TLR4 were purchased from Santa Cruz (Santa Cruz, USA). The antibodies for pNF- κB, NF-κB, NLRP3, cyclooXygenase-2 (COX-2), Iba1 and DCX were purchased from Abcam (Cambridge, USA).

2.3. Molecular docking

Crystal structures of the TLR4-MD2 complex (PDB code, 3FXI), NF- κB (PDB code, 1IKN) and NLRP3 (PDB code, 6NPY) were obtained from the RCSB Protein Data Bank (http://www.rcsb.org/). PyMOL and AutoDockTools were used to process the proteins and ligands. AutoDock Vina was used to calculate the binding energy and structure. Discovery Studio Visualizer was used to analyze the interaction sites and produce the cartoon video.

2.4. Drug administration

Seventy-two mice were randomly divided into the following 6 groups (n 12): Control-vehicle (saline), LPS-vehicle (saline), LPS-fluoXetine (20 mg/kg), LPS-Paeoniflorin (20 mg/kg), LPS-Paeoniflorin (40 mg/ kg), LPS-Paeoniflorin (80 mg/kg). Paeoniflorin was dissolved and fluoXetine was suspended in saline and administered by oral gavage once daily for 1 week. Mice were injected with LPS (0.83 mg/kg) or saline intraperitoneally (i.p.) after the last administration. A sucrose preference test was performed between 0 h and 24 h after the single LPS challenge. Forced swimming test was performed immediately after the sucrose preference test. Then the animals were put back to their cages until dry, followed by sacrifice. The timeline of the experiment was provided in Fig. 2A. The body weight of each mouse was measured after LPS injection at 0 h and 24 h. The doses of paeoniflorin were adopted according to a previous study (Li et al., 2017). LPS caused sickness symptoms such as decreased food intake and body weight as well as reduced locomotor activity. However, the sickness symptoms peak from 2 to 6 h after LPS injection and gradually disappear (Vichaya et al., 2019b). Depressive symptoms appear to peak from 6 to 24 h in this background (Dantzer et al., 2008). More importantly, LPS induced reduction of sucrose preference was observed at 24 h post LPS injection but not 48 h or 72 h (Vichaya et al., 2019a). The reduction resolves by 24 h in both young adult and aged mice (Martin et al., 2014). On the other hand, our preliminary study also showed that LPS did not affect body weight during 12–24 h and did not alter locomotor activity at 24 h post LPS injection. Therefore, the sucrose preference test was performed 0–24 h post LPS injection and the forced swimming test was performed 24 h post LPS injection.
In another independent experiment, forty mice were randomly divided into the following 5 groups (n 8): Control-vehicle groups (saline), Control-fluoXetine (20 mg/kg), Control-Paeoniflorin (20 mg/ kg), Control-Paeoniflorin (40 mg/kg), Control-Paeoniflorin (80 mg/kg). Saline was intraperitoneally injected after 1 week of administration of paeoniflorin or fluoXetine. Behavioral tests were performed as described above.

2.5. Sucrose preference test

The sucrose preference test was performed as previously described (Cheng et al., 2019). Before the experiment, sucrose training was per- formed to confirm the normal sucrose preference of the mice. Before testing, the mice were trained to adjust the sucrose solution (1%, w/v): two bottles of sucrose solution were placed in each cage for 24 h, and then a bottle of water was used to replace a bottle of sucrose solution for 24 h. After the adaptation, water was deprived for 12 h. The formal sucrose preference test was carried out between 0 h and 24 h post LPS injection. During the test, the mice were placed in separate cages and each was free to drink the sucrose solution and water in two bottles. The weight of the consumed sucrose solution and water was recorded and calculated as sucrose preference.

2.6. Forced swimming test

The forced swimming test was performed as previously described with minor modification (Bourin et al., 2004). Mice were forced to swim in a transparent glass vessel (50 cm high, 14 cm in diameter) filled with 40 cm of water at 25 2 ◦C. The immobility time was recorded when mice made no attempts to escape except the movements necessary to keep their heads above the water. The total duration of immobility was measured by Jiliang Software during the last 4 min of the 6 min testing period (Mao et al., 2008; Wang, et al., 2018b).

2.7. Protein extraction and western blotting

Tissues were homogenized in RIPA buffer and centrifuged at 1000 g for 5 min at 4 ◦C. The supernatant was then centrifuged at 12,000 g at 4◦C for 10 min and the resulting supernatant was collected. The protein concentration in the final supernatant was determined by Bradford protein assay using bovine serum albumin as a standard. The protein was separated by SDS-polyacrylamide gel electrophoresis and trans- ferred to a polyvinylidene fluoride membrane. The membrane was then blocked with 5% (w/v) skim milk powder in Tris buffered saline con- taining 0.1% Tween 20 (TBST). The membrane was incubated with the following antibodies in TBST: anti-FGF-2 (1:1000), anti-TLR4 (1:1000), anti-pNF-κB (1:1000), anti-NF-κB (1:1000), anti-NLRP3 (1:1000) and β-actin (1:5000). After incubation, the membrane was washed three times with TBST and then incubated with horseradish peroXidase- conjugated anti-rabbit (1:2000) or anti-mouse IgG secondary antibody (1:3000) in 5% skim milk powder in TBST. After washing three times with TBST, the immunocomplexes were detected with a Tanon Chem- iluminescence Imaging System and analyzed by ImageJ software. Five samples from each group were used for analysis.

2.8. ELISA

Total protein extracts from the hippocampus were also used for cytokine measurement. The levels of proinflammatory cytokines including IL-1β, IL-6 and TNF-α in the hippocampus were measured by commercial ELISA kits according to manufacturer’s instructions. Four or siX samples from each group were used for analysis.

2.9. Brain extirpation and immunofluorescence

Mice were randomly selected and injected with anesthetic and then perfused with PBS and 4% paraformaldehyde. The brain was fiXed with the same 4% paraformaldehyde for 24 h and was gradually incubated in a 10%, 20%, and 30% gradient sucrose solution; then, the brain was embedded in OCT. After that, 18 μm thick sections were cut, fiXed in 4% paraformaldehyde at 4 ◦C, permeabilized in 1% PBST, blocked with blocking buffer, and incubated with anti-Iba1 (1:200), anti-COX-2 (1:200) or anti-DCX (1:250) antibody overnight. Thsections were washed three times with PBS and then incubated with fluorescent secondary antibodies for 5 h at 45 ◦C. Then, DAPI (1:5000) was added. Finally, Iba1 and COX-2 channels in the dentate gyrus were observed under confocal microscopy (Leica TCS SP8). Images were analyzed and counted in Image-Pro Plus software. Three or four samples from each group were used for analysis.

2.10. Golgi staining

The procedure of Golgi staining was based on our previous publi- cation (Yi et al., 2020). The brain was removed and stored in Golgi CoX solution for 14 days, and then immersed in 30% sucrose solution for 3 days. A vibration knife was used to obtain a 50 μm thick coronal section of the area to be studied. The slices were then treated with ammonium hydroXide for 30 min, and then Kodak film fiXer for 30 min, rinsed with distilled water, dehydrated and fiXed with resin medium. During morphological analysis, three pyramidal neurons in the CA3 subregion of each hemisphere were selected on three slices at A/P levels (Bregma 2.06 mm, 2.46 mm, 2.80 mm approXimately), as pyramidal cells in CA3 are the most vulnerable in depressive animals induced by chronic stress and corticosterone exposure compared with CA1 and dentate gyrus (Sousa et al., 2000). All protrusions were considered spines only if they were in direct continuity with the dendritic shaft, so the numbers of branches from the dendritic tree were quantified by tracing. Two den- dritic trees per neuron were counted by an observer blind to the treat- ment. The dendritic spine density in the CA3 subregion of the hippocampus was calculated. Three samples from each group were used for analysis.

2.11. SU5402 pretreatment experiment

Animals were anesthetized by 4% chloral hydrate injection (5 ml/ kg). A guide cannula was then implanted into the lateral ventricle (0.6 mm AP, 1.5 mm ML, 1.5 mm DV) using a stereotaxic apparatus. The guide cannula was closed with a cannula dummy cap. The animals recovered for 1 week before the formal experiments and were divided into the following groups (n = 8): Control-vehicle, LPS-vehicle, LPS- SU5402, LPS-Paeoniflorin, LPS-SU5402 Paeoniflorin. SU5402 (3 μg) was microinjected into the lateral ventricle 1 h prior to paeoniflorin (40 mg/kg) oral administration. After the last drug administration on the 7th day, LPS (0.83 mg/kg) was injected intraperitoneally, followed by behavioral tests.

2.12. Data and statistical analysis

The data are expressed as the mean ± SEM and were analyzed by one-way or two-way ANOVA followed by Bonferroni’s post hoc test. A value of p < 0.05 was defined as statistically significant. 3. Results 3.1. Molecular docking analysis of paeoniflorin with TLR4/NF-κB/ NLRP3 signaling Molecular docking by AutoDock Vina demonstrated that paeoni- florin was in the pocket of the TLR4-MD2 complex, NF-κB and NLRP3 protein (Fig. 1). The simulation showed that paeoniflorin interacts with TLR4/MD-2, NF-κB and NLRP3 mainly by hydrophobic and hydrogen bonding interactions. The binding energies between paeoniflorin and the proteins were 8.0 kcal/mol, 8.3 kcal/mol and 7.7 kcal/mol, respectively, indicating the putative inhibitory activity of paeoniflorin on inflammation. In addition, the cartoon videos regarding the receptor- ligand interaction were shown in supplementary materials (Videos S1–S3). 3.2. The depressive behaviors, the reduction of FGF-2 and the elevation of proinflammatory cytokines were reversed by paeoniflorin Table 1 shows the body weight gain in response to LPS and/or paeoniflorin treatment. During 0–24 h post LPS injection, all the groups exhibited decreased body weight gain, while control animals receiving saline obtained approXimately 10% body weight gain. Water depriva- tion 12 h prior to body weight measurement (the start of the sucrose preference test) could account for this increase. During water depriva- tion, food intake decreased, and thus, body weight should be reduced between —12 h and 0 h. This caused the lower baseline value. When the deprivation was removed (0–24 h, during the sucrose preference test), the water intake and food intake in Control animals increased sharply, so the body weight was rapidly restored. One-way ANOVA indicated a significant treatment effect on sucrose preference [F(5,66) 2.87, p < 0.05] or immobility time [F(5,66) 8.31, p < 0.01]. Post hoc test showed that LPS injection caused depressive behaviors, which were reflected by decreased sucrose preference [p < 0.01, Fig. 2B] and increased immobility time [p < 0.01, Fig. 2C]. Pretreatment with both fluoXetine (20 mg/kg) and paeoniflorin (20, 40 mg/kg) prevented the decrease in sucrose preference [p < 0.05, p < 0.05, p < 0.05] and the increase in immobility time [p < 0.01, p < 0.01, p < 0.01] in mice in response to LPS injection. In addition, paeo- niflorin at 80 mg/kg only decreased the immobility time but did not affect the sucrose preference. In addition, One-way ANOVA indicated a significant treatment effect on FGF-2 levels [F(5,24) 6.37, p < 0.01]. Post hoc test showed that LPS injection caused a significant reduction in FGF-2 levels in the hippo- campus [p < 0.01]. Pretreatment with both fluoXetine (20 mg/kg) and paeoniflorin (20, 40, 80 mg/kg) prevented the decrease in hippocampal FGF-2 levels [p < 0.01, p < 0.01, p < 0.01, p < 0.05] in mice in response to LPS injection (Fig. 2D). Moreover, One-way ANOVA indicated a significant treatment effect on IL-1β, IL-6 or TNF-α levels [F(5,30) = 3.91, p < 0.01; F(5,30) = 4.84, p < 0.01; F(5,30) = 7.22, p < 0.01]. Post hoc test showed that LPS injection caused a significant elevation of IL-1β, IL-6 and TNF-α levels in the hippocampus [p < 0.01, p < 0.01, p < 0.01, respectively]. showed that LPS injection caused an increase in TLR4 levels, NF-κB phosphorylation and NLRP3 levels in the hippocampus [p < 0.01, p < 0.01, p < 0.01, respectively]. Pretreatment with paeoniflorin prevented the upregulation of TLR4 levels, NF-κB phosphorylation and NLRP3 levels in LPS treated mice. On the other hand, paeoniflorin did not cause Pretreatment with both fluoXetine (20 mg/kg) and paeoniflorin (40, 80 mg/kg) prevented the elevation of proinflammatory cytokines in mice in response to LPS injection (Fig. 2E–G). Paeoniflorin at 20 mg/kg only decreased IL-6 and TNF-α levels. On the other hand, paeoniflorin administration ranging from 20 to 80 mg/kg did not cause any alteration in behaviors, FGF-2 and proin- flammatory cytokine levels in normal mice compared with that in any alteration in TLR4/NF-κB/NLRP3 levels in normal animals (Fig. S1). 3.4. The effects of paeoniflorin on hippocampal Iba1/COX-2 in mice One-way ANOVA indicated a significant treatment effect on Iba1 positive cells [F(5,12) 20.43, p < 0.01]. Post hoc test showed that LPS injection caused a significant elevation of Iba1 labeled microglia in the dentate gyrus [p < 0.01]. Pretreatment with both fluoXetine (20 mg/kg) and paeoniflorin (20, 40, 80 mg/kg) prevented the elevation of Iba1 positive cells [p < 0.01, p < 0.01, p < 0.01, p < 0.01, respectively] in mice in response to LPS injection (Fig. 4). In addition, One-way ANOVA indicated a significant treatment effect on COX-2 levels [F(5,12) 96.71, p < 0.01]. Post hoc test showed that LPS injection also caused a significant increase in the fluorescence intensity of COX-2 in the dentate gyrus [p < 0.01]. Pretreatment with both fluoXetine (20 mg/kg) and paeoniflorin (20, 40, 80 mg/kg) prevented the increase in COX-2 in- tensity [p < 0.01, p < 0.01, p < 0.01, p < 0.01, respectively] in mice in response to LPS injection (Fig. 4). 3.5. The effects of paeoniflorin on the dendritic spine in the hippocampus One-way ANOVA indicated a significant treatment effect on den- dritic spine density [F(5,12) 6.24, p < 0.01]. Post hoc test showed that LPS injection caused a significant inhibition of dendritic spine density in the CA3 region of the hippocampus [p < 0.05]. Pretreatment with both fluoXetine (20 mg/kg) and paeoniflorin (only at 40 mg/kg) prevented the inhibition of dendritic spine density [p < 0.05, p < 0.05, respec- tively] in mice in response to LPS injection (Fig. 5). 3.6. FGF-2 levels in hippocampal neurons were increased by paeoniflorin in response to LPS In addition to the results from Western blot, double staining immu- nofluorescence was also used for FGF-2 verification. One-way ANOVA indicated a significant treatment effect on FGF-2 intensity [F(5,12) 128.70, p < 0.01]. Post hoc test showed that LPS injection resulted in a significant decrease of FGF-2 fluorescence intensity in NeuN labeled neurons in the dentate gyrus [p < 0.01], while this decrease was prevented [p < 0.05, p < 0.05, p < 0.01, p < 0.01] by the pretreatment with fluoXetine (20 mg/kg) and paeoniflorin (20, 40, 80 mg/kg) in mice response to LPS injection (Fig. 6). 3.7. FGFR1 inhibitor SU5402 prevented the effects of paeoniflorin on behaviors and proinflammatory cytokines As shown in Fig. 7A and B, Two-way ANOVA indicated that a sig- nificant pretreatment effect [F(1,28) 11.49, p < 0.01; F(1,28) 16.21, p < 0.01], a nonsignificant treatment effect and the significant inter- action [F(1,28) 9.32, p < 0.01; F(1,28) 27.13, p < 0.01]on sucrose preference and immobility time. LPS injection decreased sucrose pref- erence [p < 0.01] and increased immobility time [p < 0.01] in mice, while paeoniflorin increased sucrose preference [p < 0.05] and decreased immobility time [p < 0.01]. However, the effects of paeoni- florin were prevented by pretreatment with SU5402 [p < 0.01, p < 0.01, respectively]. In addition, there was no significant effect or interaction on proinflammatory cytokines according to a Two-way ANOVA. LPS increased IL-1β [p < 0.05] and IL-6 levels [p < 0.05] (Fig. 7C and D), while paeoniflorin decreased IL-1β levels [p < 0.05]. 3.8. FGFR1 inhibitor SU5402 prevented the effects of paeoniflorin on Iba1/COX-2 intensity in the hippocampus As shown in Fig. 8, Two-way ANOVA indicated that a pretreatment effect [F(1,12) 54.27, p < 0.01; F(1,12) 41.48, p < 0.01], a treatment effect [F(1,12) 75.21, p < 0.01; F(1,12) 1.02, p > 0.05] and the interaction [F(1,12) 17.61, p < 0.01; F(1,12) 22.40, p < 0.01] on Iba1 positive cells and COX-2 intensity. LPS injection increased the number of Iba1 positive cells [p < 0.01] and the intensity of COX-2 in the hippocampus [p < 0.01], while paeoniflorin decreased the number [p < 0.05] and intensity [p < 0.05]. However, the effects of paeoniflorin were prevented by pretreatment with SU5402 [p < 0.05, p < 0.01, respectively]. 3.9. FGFR1 inhibitor SU5402 prevented the neurogenic effects of paeoniflorin in the hippocampus As shown in Fig. 9, Two-way ANOVA indicated that a significant pretreatment effect [F(1,12) 11.32, p < 0.01], a significant treatment effect [F(1,12) 9.17, p < 0.05] and the nonsignificant interaction [F (1,12) 1.81, p > 0.05] on DCX positive cells. LPS injection inhibited the neurogenesis in the hippocampus as the number of DCX positive cells was decreased [p < 0.05]. The number of DCX positive cells was increased by paeoniflorin [p < 0.05] but prevented by SU5402 [p < 0.05]. 4. Discussion In the current study, the neuroinflammation/FGF-2 related mecha- nism of paeoniflorin was investigated in LPS induced animal model of depression. The results showed that paeoniflorin exhibited effects in both the sucrose preference test and forced swimming test. There is growing evidence suggesting that inflammation plays a role in the pathophysiology of depression (Felger, 2018). To investigate how paeoniflorin exerts its action in LPS induced depression, molecular docking was firstly used for simulation. TLR4/NF-κB/NLRP3 signaling has been shown to regulate inflammatory responses in microglia (Zhong et al., 2019). Activation of TLR4/NF-κB/NLRP3 promotes the over- expression and overrelease of proinflammatory cytokines, which are widely known to cause neuronal injury (Hill et al., 2019; Jiang et al., 2019). The docking showed that the binding energies between paeoni- florin and TLR4/NF-κB/NLRP3 were less than 7 kcal/mol, indicating good bonding performance. According to the docking structure, paeo- niflorin competitively binds with the active sites of TLR4/MD-2 and NLRP3 by hydrophobic and hydrogen bonding interactions, which block the interaction between agonist LPS/NEK7 and its target TLR4/NLRP3. In accordance with the docking results, western blotting showed that LPS caused the upregulation of TLR4 levels, NF-κB phosphorylation and NLRP3 levels, while paeoniflorin prevented the abnormalities. Consis- tently, the results also showed that LPS caused a significant elevation in IL-1β, IL-6 and TNF-α levels. In contrast, the proinflammatory cytokines were decreased by paeoniflorin treatment. Similarly, the classic anti- depressant fluoXetine also decreased proinflammatory cytokines in the hippocampus, which was in parallel to a previous study showing that fluoXetine reduced the levels of inflammatory factors in the central nervous system (Tynan et al., 2012). These observations clearly sug- gested that anti-inflammatory activity was involved in the antidepres- sant effects of paeoniflorin. It is well known that microglia are recognized as protective to neu- rons when they are in a quiescent state. However, the effects of microglia on neurons are harmful when microglia are classically activated (Takahashi et al., 2005). The present study results showed that LPS caused the overexpression of Iba1, a microglial marker in the hippo- campus, indicating microglial activation. However, this number of Iba1 positive cells was alleviated by paeoniflorin treatment, demonstrating the inactivation of microglia, which was partly consistent with previous reports showing that paeoniflorin inhibited microglial activation in neuropathic pain and inflammatory pain (Hu et al., 2018; Li et al., 2017; Zhou et al., 2019). On the other hand, COX derived oXidative stress is one of the biomarkers in inflammation (Chung et al., 2002), and COX-2 is associated with inflammatory stimuli such as microglial activation (Liu et al., 2017). Moreover, the current study suggests a detrimental rational to hypothesize that FGF-2 mediates the effects of antidepres- sants. In this way, we found that LPS decreased the levels of FGF-2 in the hippocampus, which was in accordance to the recent study showing neuroinflammation caused the reduction of FGF-2 (Chen et al., 2020). There is evidence that microglia play an important role in synaptic Administration of both fluoXetine and paeoniflorin significantly pruning(Schafer et al., 2012). Microglia affect and eliminate synapses. If the microglia numbers are reduced or microglia lack activity in the brain, the development of the brain and synaptic pruning will be delayed (Paolicelli et al., 2011). This deficiency in synaptic pruning will subse- quently result in excess of dendritic spines. However, if the microglial activity is excessively activated, the process of synaptic pruning will be harmful to the normal function of the central nervous system (Sellgren et al., 2019). Modulating microglia mediated synaptic pruning is likely to lead to controlling synaptic homeostasis and the interactions between the brain and immune system (Marinelli et al., 2019). In parallel to the change in microglia number, the dendritic spine density was decreased by LPS injection but reversed by paeoniflorin, indicating that synaptic pruning returns to a normal state. FGF-2 is considered a putative target for antidepressants, as endog- enous FGF-2 injection directly produces antidepressant effects in ani- mals (Elsayed et al., 2012; Simard et al., 2018; Wang, et al., 2018a). Considering that FGF-2 participates in regulating neurogenesis, it is reversed this abnormality, indicating that paeoniflorin reversed the reduction of FGF-2 caused by neuroinflammation. Furthermore, to assess whether the recovery of neuroinflammation induced FGF-2 reduction was necessary for the antidepressant effects of paeoniflorin, an intervention experiment was performed by pretreatment with the FGFR1 inhibitor SU5402. The results showed that SU5402 prevented the antidepressant effect of paeoniflorin as SU5402 ameliorated the improvement of anhedonia and despair behaviors induced by paeoni- florin. More importantly, paeoniflorin administration alone inhibited the activation of microglia and the enhancement of COX-2 intensity, however, SU5402 blocked these improvements, indicating that the anti-inflammatory activity of paeoniflorin required the activation of FGF-2/FGFR1 signaling. In parallel to the microglial activation changes, the release of IL-6 and TNF-αinhibited by paeoniflorin was reversed after coadministration with SU5402. The results suggested that the inacti- vation of FGF-2 signaling not only blocked the neurogenesis, but also prevented the effects of paeoniflorin on neuroinflammation. This could account for why FGFR antagonist could block the antidepressant effects of paeoniflorin. Notably, SU5402 could not block the effects of paeoni- florin on IL-1β release, although microglia were reactivated. One pos- sibility for the downregulation of IL-1β despite microglial activation is the binding interaction between paeoniflorin and NLPR3, which blocks the processing of mature IL-1β by NLRP3. On the other hand, it should be noted that paeoniflorin and fluoXe- tine did not significantly decrease the immobility time of normal ani- mals in the forced swimming test. These results were not consistent with a previous study (Qiu, et al., 2013a), showing that paeoniflorin at 40 mg/kg decreased the immobility time in both the forced swimming and tail suspension tests. We speculate that the time windows between drug administration and behavioral test might affect the results, as the forced swimming test and tail suspension test were performed 1 h after paeo- niflorin administration in the previous study. However, the forced swimming test was performed 24 h after paeoniflorin administration in the present study. Similarly, antidepressants ketamine and fluoXetine administered 1 h before forced swimming test or tail suspension test decreased immobility time in normal animals (Takahashi et al., 2020; Yi et al., 2010). On the contrary, ketamine and fluoXetine administered 24 h before forced swimming test did not decrease immobility time in normal animals, but decreased immobility time in LPS treated animals (Taniguti et al., 2019; Walker et al., 2013). Taken together, activation of microglia causes excessive synaptic pruning in neurons. The injured neurons could not secrete sufficient FGF-2 to reduce the hyperactivity of microglia, which further aggravates neuroinflammation. These observations implied that paeoniflorin reduced the effects of LPS by binding with TLR4/NF-κB/NLRP3, decreased the release of proinflammatory cytokines and prevented neuronal impairment from cytokines. In this respect, neurons were rescued from secreting FGF-2, which further inhibited the activation of microglia and neuroinflammation. References Bourin, M., Mocaer, E., Porsolt, R., 2004. Antidepressant-like activity of S 20098 (agomelatine) in the forced swimming test in rodents: involvement of melatonin and serotonin receptors. J. Psychiatry Neurosci. 29 (2), 126–133. Buckwalter, M.S., Wyss-Coray, T., 2004. Modelling neuroinflammatory phenotypes in vivo. J. Neuroinflammation 1 (1), 10. Casaril, A.M., Domingues, M., de Andrade Lourenco, D., Birmann, P.T., Padilha, N., Vieira, B., Begnini, K., SeiXas, F.K., Collares, T., Lenardao, E.J., Savegnago, L., 2019. Depression- and anxiogenic-like behaviors induced by lipopolysaccharide in mice are reversed by a selenium-containing indolyl compound: behavioral, neurochemical and computational insights involving the serotonergic system. J. Psychiatr. Res. 115, 1–12. Catena-Dell’Osso, M., Bellantuono, C., Consoli, G., Baroni, S., Rotella, F., Marazziti, D., 2011. Inflammatory and neurodegenerative pathways in depression: a new avenue for antidepressant development? Curr. Med. Chem. 18 (2), 245–255. Chen, L.B., Qiu, F.M., Zhong, X.M., Hong, C., Huang, Z., 2019. Promoting neurogenesis in hippocampal dentate gyrus of chronic unpredictable stress-induced depressive-like rats with paeoniflorin. J. Integr. Neurosci. 18 (1), 43–49. Chen, M., Zhang, Q.P., Zhu, J.X., Cheng, J., Liu, Q., Xu, G.H., Li, C.F., Yi, L.T., 2020. Involvement of FGF-2 modulation in the antidepressant-like effects of liquiritin in mice. Eur. J. Pharmacol. 881, 173297. Chen, X.Q., Li, C.F., Chen, S.J., Liang, W.N., Wang, M., Wang, S.S., Dong, S.Q., Yi, L.T., Li, C.D., 2018. The antidepressant-like effects of Chaihu Shugan San: dependent on the hippocampal BDNF-TrkB-ERK/Akt signaling activation in perimenopausal depression-like rats. Biomed. Pharmacother. 105, 45–52. Cheng, J., Chen, M., Zhu, J.X., Li, C.F., Zhang, Q.P., Geng, D., Liu, Q., Yi, L.T., 2019. FGF- 2 signaling activation in the hippocampus contributes to the behavioral and cellular responses to puerarin. Biochem. Pharmacol. 168, 91–99. Chung, H.Y., Kim, H.J., Kim, K.W., Choi, J.S., Yu, B.P., 2002. Molecular inflammation NSC 178886 hypothesis of aging based on the anti-aging mechanism of calorie restriction. Microsc. Res. Tech. 59 (4), 264–272.
Dantzer, R., O’Connor, J.C., Freund, G.G., Johnson, R.W., Kelley, K.W., 2008. From inflammation to sickness and depression: when the immune system subjugates the brain. Nat. Rev. Neurosci. 9 (1), 46–56.
Elsayed, M., Banasr, M., Duric, V., Fournier, N.M., Licznerski, P., Duman, R.S., 2012. Antidepressant effects of fibroblast growth factor-2 in behavioral and cellular models of depression. Biol Psychiatry 72 (4), 258–265.
Enache, D., Pariante, C.M., Mondelli, V., 2019. Markers of central inflammation in major depressive disorder: a systematic review and meta-analysis of studies examining cerebrospinal fluid, positron emission tomography and post-mortem brain tissue. Brain Behav. Immun. 81, 24–40.
Felger, J.C., 2018. Imaging the role of inflammation in mood and anxiety-related disorders. Curr. Neuropharmacol. 16 (5), 533–558.
Guo, R.B., Wang, G.F., Zhao, A.P., Gu, J., Sun, X.L., Hu, G., 2012. Paeoniflorin protects against ischemia-induced brain damages in rats via inhibiting MAPKs/NF-kappaB- mediated inflammatory responses. PLoS One 7 (11) e49701.
Hill, J.D., Zuluaga-Ramirez, V., Gajghate, S., Winfield, M., Sriram, U., Rom, S., Persidsky, Y., 2019. Activation of GPR55 induces neuroprotection of hippocampal neurogenesis and immune responses of neural stem cells following chronic, systemic inflammation. Brain. Behav. Immun. 76, 165–181.
Hu, B., Xu, G., Zhang, X., Xu, L., Zhou, H., Ma, Z., Shen, X., Zhu, J., Shen, R., 2018. Paeoniflorin attenuates inflammatory pain by inhibiting microglial activation and akt-NF-kappaB signaling in the central nervous system. Cell. Physiol. Biochem. 47 (2), 842–850.
Hu, M.Z., Wang, A.R., Zhao, Z.Y., Chen, X.Y., Li, Y.B., Liu, B., 2019. Antidepressant-like effects of paeoniflorin on post-stroke depression in a rat model. Neurol. Res. 41 (5), 446–455.
Jesulola, E., Micalos, P., Baguley, I.J., 2018. Understanding the pathophysiology of depression: from monoamines to the neurogenesis hypothesis model – are we there yet? Behav. Brain Res. 341, 79–90.
Jiang, J., Yu, Y., Kinjo, E.R., Du, Y., Nguyen, H.P., Dingledine, R., 2019. Suppressing proinflammatory prostaglandin signaling attenuates excitotoXicity-associated neuronal inflammation and injury. Neuropharmacology 149, 149–160.
Katsouri, L., Ashraf, A., Birch, A.M., Lee, K.K., Mirzaei, N., Sastre, M., 2015. Systemic administration of fibroblast growth factor-2 (FGF2) reduces BACE1 expression and amyloid pathology in APP23 mice. Neurobiol. Aging 36 (2), 821–831.
Li, J., Huang, S., Huang, W., Wang, W., Wen, G., Gao, L., Fu, X., Wang, M., Liang, W., Kwan, H.Y., Zhao, X., Lv, Z., 2017. Paeoniflorin ameliorates interferon-alpha- induced neuroinflammation and depressive-like behaviors in mice. Oncotarget 8 (5), 8264–8282.
Liu, N., Zhuang, Y., Zhou, Z., Zhao, J., Chen, Q., Zheng, J., 2017. NF-kappaB dependent up-regulation of TRPC6 by Abeta in BV-2 microglia cells increases COX-2 expression and contributes to hippocampus neuron damage. Neurosci. Lett. 651, 1–8.
Liu, S.C., Hu, W.Y., Zhang, W.Y., Yang, L., Li, Y., Xiao, Z.C., Zhang, M., He, Z.Y., 2019a. Paeoniflorin attenuates impairment of spatial learning and hippocampal long-term potentiation in mice subjected to chronic unpredictable mild stress. Psychopharmacology (Berl) 236 (9), 2823–2834.
Liu, X., Zheng, X., Du, G., Li, Z., Qin, X., 2019b. Brain metabonomics study of the antidepressant-like effect of Xiaoyaosan on the CUMS-depression rats by (1)H NMR analysis. J. Ethnopharmacol. 235, 141–154.
Mao, Q.Q., Ip, S.P., Tsai, S.H., Che, C.T., 2008. Antidepressant-like effect of peony glycosides in mice. J. Ethnopharmacol. 119 (2), 272–275.
Marinelli, S., Basilico, B., Marrone, M.C., Ragozzino, D., 2019. Microglia-neuron crosstalk: signaling mechanism and control of synaptic transmission. Semin. Cell Dev. Biol. 94, 138–151.
Martin, S.A., Dantzer, R., Kelley, K.W., Woods, J.A., 2014. Voluntary wheel running does not affect lipopolysaccharide-induced depressive-like behavior in young adult and aged mice. Neuroimmunomodulation 21 (1), 52–63.
Paolicelli, R.C., Bolasco, G., Pagani, F., Maggi, L., Scianni, M., Panzanelli, P., Giustetto, M., Ferreira, T.A., Guiducci, E., Dumas, L., Ragozzino, D., Gross, C.T., 2011. Synaptic pruning by microglia is necessary for normal brain development. Science 333 (6048), 1456–1458.
Phan, T.X., Malkani, R.G., 2019. Sleep and circadian rhythm disruption and stress intersect in Alzheimer’s disease. Neurobiol. Stress 10, 100133.
Polazzi, E., Contestabile, A., 2002. Reciprocal interactions between microglia and neurons: from survival to neuropathology. Rev. Neurosci. 13 (3), 221–242.
Qiu, F., Zhong, X., Mao, Q., Huang, Z., 2013a. The antidepressant-like effects of paeoniflorin in mouse models. EXp. Ther. Med. 5 (4), 1113–1116.
Qiu, F.M., Zhong, X.M., Mao, Q.Q., Huang, Z., 2013b. Antidepressant-like effects of paeoniflorin on the behavioural, biochemical, and neurochemical patterns of rats exposed to chronic unpredictable stress. Neurosci. Lett. 541, 209–213.
Schafer, D.P., Lehrman, E.K., Kautzman, A.G., Koyama, R., Mardinly, A.R., Yamasaki, R., Ransohoff, R.M., Greenberg, M.E., Barres, B.A., Stevens, B., 2012. Microglia sculpt postnatal neural circuits in an activity and complement-dependent manner. Neuron 74 (4), 691–705.
Sellgren, C.M., Gracias, J., Watmuff, B., Biag, J.D., Thanos, J.M., Whittredge, P.B., Fu, T., Worringer, K., Brown, H.E., Wang, J., Kaykas, A., Karmacharya, R., Goold, C.P., Sheridan, S.D., Perlis, R.H., 2019. Increased synapse elimination by microglia in schizophrenia patient-derived models of synaptic pruning. Nat. Neurosci. 22 (3), 374–385.
Simard, S., Shail, P., MacGregor, J., El Sayed, M., Duman, R.S., Vaccarino, F.M., Salmaso, N., 2018. Fibroblast growth factor 2 is necessary for the antidepressant effects of fluoXetine. PloS One 13 (10) e0204980.
Sousa, N., Lukoyanov, N.V., Madeira, M.D., Almeida, O.F., Paula-Barbosa, M.M., 2000. Reorganization of the morphology of hippocampal neurites and synapses after stress- induced damage correlates with behavioral improvement. Neuroscience 97 (2), 253–266.
Tajfard, M., Latiff, L.A., Rahimi, H.R., Mouhebati, M., Esmaeily, H., Taghipour, A., Mahdipour, E., Davari, H., Saghiri, Z., Hanachi, P., Ghayour Mobarhan, M., Ferns, G. A., Azizian, M., 2014. Serum inflammatory cytokines and depression in coronary artery disease. Iran. Red Crescent Med. J. 16 (7) e17111.
Takahashi, K., Kitamura, Y., Ushio, S., Sendo, T., 2020. Immobility-reducing effects of ketamine during the forced swim test on 5-HT1A receptor activity in the medial prefrontal cortex in an intractable depression model. Acta Med. Okayama 74 (4), 301–306.
Takahashi, K., Rochford, C.D., Neumann, H., 2005. Clearance of apoptotic neurons without inflammation by microglial triggering receptor expressed on myeloid cells- 2. J. EXp. Med. 201 (4), 647–657.
Tang, M.M., Lin, W.J., Pan, Y.Q., Li, Y.C., 2018. Fibroblast growth factor 2 modulates hippocampal microglia activation in a neuroinflammation induced model of depression. Front. Cell. Neurosci. 12, 255.
Taniguti, E.H., Ferreira, Y.S., Stupp, I.J.V., Fraga-Junior, E.B., Doneda, D.L., Lopes, L., Rios-Santos, F., Lima, E., Buss, Z.S., Viola, G.G., Vandresen-Filho, S., 2019. Atorvastatin prevents lipopolysaccharide-induced depressive-like behaviour in mice. Brain Res. Bull. 146, 279–286.
Tian, D.D., Wang, M., Liu, A., Gao, M.R., Qiu, C., Yu, W., Wang, W.J., Zhang, K., Yang, L., Jia, Y.Y., Yang, C.B., Wu, Y.M., 2021. Antidepressant effect of paeoniflorin is through inhibiting pyroptosis CASP-11/GSDMD pathway. Mol. Neurobiol. 58 (2), 761–776.
Tynan, R.J., Weidenhofer, J., Hinwood, M., Cairns, M.J., Day, T.A., Walker, F.R., 2012. A comparative examination of the anti-inflammatory effects of SSRI and SNRI antidepressants on LPS stimulated microglia. Brain Behav. Immun. 26 (3), 469–479.
Vichaya, E.G., Gross, P.S., Estrada, D.J., Cole, S.W., Grossberg, A.J., Evans, S.E., Tuvim, M.J., Dickey, B.F., Dantzer, R., 2019a. Lipocalin-2 is dispensable in inflammation-induced sickness and depression-like behavior. Psychopharmacology (Berl) 236 (10), 2975–2982.
Vichaya, E.G., Laumet, G., Christian, D.L., Grossberg, A.J., Estrada, D.J., Heijnen, C.J., Kavelaars, A., Dantzer, R., 2019b. Motivational changes that develop in a mouse model of inflammation-induced depression are independent of indoleamine 2,3 dioXygenase. Neuropsychopharmacology 44 (2), 364–371.
Walker, A.K., Budac, D.P., Bisulco, S., Lee, A.W., Smith, R.A., Beenders, B., Kelley, K.W., Dantzer, R., 2013. NMDA receptor blockade by ketamine abrogates lipopolysaccharide-induced depressive-like behavior in C57BL/6J mice. Neuropsychopharmacology 38 (9), 1609–1616.
Wang, L., Li, X.X., Chen, X., Qin, X.Y., Kardami, E., Cheng, Y., 2018a. Antidepressant-like effects of low- and high-molecular weight FGF-2 on chronic unpredictable mild stress mice. Front. Mol. Neurosci. 11, 377.
Wang, S., Zhao, X., Qiao, Z., Jia, X., Qi, Y., 2018b. Paeoniflorin attenuates depressive behaviors in systemic lupus erythematosus mice. Biomed. Pharmacother. 103, 248–252.
Wyss-Coray, T., Mucke, L., 2002. Inflammation in neurodegenerative disease–a double- edged sword. Neuron 35 (3), 419–432.
Yi, L.T., Dong, S.Q., Wang, S.S., Chen, M., Li, C.F., Geng, D., Zhu, J.X., Liu, Q., Cheng, J., 2020. Curcumin attenuates cognitive impairment by enhancing autophagy in chemotherapy. Neurobiol. Dis. 136, 104715.
Yi, L.T., Li, C.F., Zhan, X., Cui, C.C., Xiao, F., Zhou, L.P., Xie, Y., 2010. Involvement of monoaminergic system in the antidepressant-like effect of the flavonoid naringenin in mice. Prog. Neuro-Psychopharmacol. Biol. Psychiatry 34 (7), 1223–1228.
Zhong, X., Li, G., Qiu, F., Huang, Z., 2018. Paeoniflorin ameliorates chronic stress- induced depression-like behaviors and neuronal damages in rats via activation of the ERK-CREB pathway. Front. Psychiatry 9, 772.
Zhong, X., Liu, M., Yao, W., Du, K., He, M., Jin, X., Jiao, L., Ma, G., Wei, B., Wei, M., 2019. Epigallocatechin-3-Gallate attenuates microglial inflammation and neurotoXicity by suppressing the activation of canonical and noncanonical inflammasome via TLR4/NF-kappaB pathway. Mol. Nutr. Food Res. 63 (21) e1801230.
Zhou, D., Zhang, S., Hu, L., Gu, Y.F., Cai, Y., Wu, D., Liu, W.T., Jiang, C.Y., Kong, X., Zhang, G.Q., 2019. Inhibition of apoptosis signal-regulating kinase by paeoniflorin attenuates neuroinflammation and ameliorates neuropathic pain. J. Neuroinflammation 16 (1), 83.