Microglia in depression: an overview of microglia in the pathogenesis and treatment of depression

During brain development, microglia regulate synaptic transmission, prune neuronal synapses, and facilitate the formation of neural circuits. Once homeostasis is disturbed, microglia are converted into an active state and release pro-inflammatory cytokines, chemokines, and reactive oxidants. The pro-inflammatory microglia are involved with the upregulation of pro-inflammatory mediators, which is usually considered a harmful event. Whereas anti-inflammatory microglia display protective effects in neuronal survival and adult neurogenesis [91].

Activation of microglia is a critical mechanism in neurogenesis inhibition in the presence of inflammation and stress [9]. Hippocampal neurogenesis is a complex neurobiological process involving the generation and functional integration of newborn cells into brain neural circuits [92]. Several studies have demonstrated that stress strongly suppresses adult hippocampal neurogenesis. Both acute and chronic stress have been reported to reduce adult neurogenesis by decreasing neuroprogenitor proliferation and newborn cell survival [93‐95], and also impair newborn neuron maturation [96, 97]. LPS infusion via a cannula for 4 weeks reduced the survival of new neurons in the subgranular zone/granule cell layer (SGZ/GCL) of the HIP, while 6 days of intracortical LPS infusion in rats did not affect the proliferation of new cells [98]. In addition, LPS-induced neuroinflammation suppressed proliferation and differentiation of neural stem cells in the dentate gyrus (DG) of the HIP, which was indicated by the decreased number of BrdU-, DCX- and NeuN-positive cells [99]. In an 8-week CMS-induced depression paradigm, C57BL/6 mice exhibited an increase in the number of Iba-1 positive cells and levels of IL-1β, IL-6, and TNF-α [100]. Another report on the CMS showed that an imbalance of peripheral inflammation markers and decreased CX3CL1/CX3CR1 immunoreactivity were associated with reduced numbers of BrdU/Ki-67/DCX⁺ (nascent, proliferating and DCX-associated) cells in the DG of the HIP [101]. Furthermore, in a graded study of the CMS for 5, 6, or 7 weeks, the investigators found that the pro- or anti-inflammatory microglial changes in the cortex and HIP, as well as BrdU/DCX⁺ cells, decreased in the DG [102]. Similar results related to the relationship between depressive-like behaviors and changes in microglia and neurogenesis have been reported in psychosocial stress models, including maternal separation [103], maternal sleep deprivation [104], chronic water-immersion restraint stress (CWIRS) [105], and CSDS [106, 107]. On the other hand, IL-4 and IL-10 induce alternative activation of anti-inflammatory factors that play a critical neuroprotective function in tissue remodeling and nerve regeneration. Treatment with IL-4 markedly inhibited IL-1β-caused depressive behavior by regulating glial activation and neurotransmitter levels in the HIP [108]. IL-4 stimulation increased the expression of insulin-like growth factor-1 in microglia, which has been reported to promote neurogenesis [109, 110]. Recent research demonstrated that IL-4-driven Arg1⁺ microglia modulate stress resilience through brain-derived neurotrophic factor (BDNF)-dependent neurogenesis in CMS mice [111]. IL-10-stimulated microglia enhanced the proliferation of NPCs but did not affect neuronal differentiation. IL-10-secreting microglia supported neuronal differentiation and the survival of newly formed cells in vitro [112]. CX3CR1^CreERIL-10 knockout mice exhibited depression- and anxiety-like behaviors, along with decreased NR2B (N-methyl-D-aspartate receptor (NMDAR) subunit) and synaptophysin (SYP) levels in the mPFC, and increased NR2B and postsynaptic densitin-95 (PSD95) levels in the amygdala [113]. Furthermore, hippocampal neurogenesis was reduced in CX3CR1-deficient mice, and antagonizing CX3CR1 resulted in increased hippocampal IL-1β level and decreased neurogenesis in young rats [114]. Minocycline, a microglial inhibitor, reversed the pathogenic phagocytic potential of neurotoxic microglia, and reduced the negative phenotypes associated with reduced neurogenesis in depression models induced by chronic stress and LPS [115]. Therefore, these data suggest that microglia are correlated with reduced neurogenesis induced by exposure to stress and, in turn, may be responsible for the development of depressive-like behaviors.

The NLRP3 inflammasome is a multi-molecule complex containing cytosolic NLRP3, adaptor protein ASC, and pro-caspase-1 precursor. NLRP3 plays a vital role in production of pro-inflammatory cytokines during the stress process [116]. The NLRP3 inflammasome is currently considered to be an essential molecular platform for regulating pro-inflammatory cytokines release. NF-κB nuclear translocation is an indispensable event in the initiation of NLRP3 activation, which further confirmed that NF-κB nuclear translocation might play a role in assembling the NLRP3 inflammasome. When NLRP3 is activated by repetitive stress, it modulates caspase-1 activation, which, in turn, promotes IL-1β and IL-18 maturation in microglia, where excessive secretion of cytokines contributes to the development and progression of MDD [117‐119].

NLRP3 inflammasome activation has even been observed in depressive patients and numerous animal models of depression. Previous studies have revealed that the NLRP3 inflammasome was activated in blood cells from MDD patients, and the levels of IL-1β and IL-18 were increased in serum. The increase in pro-inflammatory factors was positively correlated with Beck Depression Inventory scores [120], suggesting that NLRP3 might have a critical role in mediating the development of depression [121]. In another experimental study, a sustained CMS procedure (12 weeks) enhanced the levels of IL-1, NLRP3, ASC, TLR2, NF-κB, p-IKKα, and IKKβ in rat PFC, while these alterations were reversed with fluoxetine [122]. Immunofluorescent analysis has confirmed that the NLRP3 inflammasome was primarily activated only in microglia in the PFC (Iba-1/NeuN⁺ cells) [122]. Using NLRP3 inflammasome inhibitors AC-YVAD-CMK and VX-765 significantly improved depressive-like behavior, as shown by improvements in sucrose intake in the sucrose preference test (SPT) and immobility time in the FST of CMS mice [94, 123]. Moreover, activation of P2X7R and NLRP3 inflammasome-associated proteins in hippocampal microglia could mediate depressive-like behaviors, which provide new therapeutic targets for depression [124]. In a LPS-induced acute depression mouse model, NLRP3, ASC, and caspase-1 mRNA expressions were remarkably increased compared to the control group [125]. In addition, IL-1β secretion was closely controlled by the NLRP3 inflammasome, which plays an important role in the pathogenesis of depression [108]. Minocycline has an acute antidepressant role in CRS, CUMS, and LH models by inhibiting microglia and NLRP3 activation [126‐129]. Fluoxetine conferred an antidepressant effect in part by inhibiting NLRP3 inflammasome activation [130]. Clomipramine reversed LPS-induced increases in IL-1β, IL-6, TNF-α, and NLRP3 gene expressions in vivo, as well as in vitro in BV2 cells [131]. Electro-acupuncture treatment for 4 weeks markedly reversed CMS-induced increases in NLRP3 inflammasome-associated components (NLRP3, ASC, and caspase-1) and the expressions of mature IL-1β, as well as IL-18, TNF-α, IL-6, P2X7 receptor, and Iba-1 [132]. Iptakalim negatively regulates NLRP3 expression and, in turn, affects microglia-mediated neuroinflammation by inhibiting the activation of NLRP3/caspase-1 axis in the HIP of CMS mice [94]. These findings suggest that microglial NLRP3 activation is a central mediator involved in depressive-like behaviors in animal models and MDD patients.

Chronic stress stimulation or inflammation may contribute to tryptophan (TRP) metabolism associated with the kynurenine pathway (KP). The generation of neuroactive kynurenine metabolites leads to subsequent depressive-like symptoms [133]. KP metabolizes tryptophan into several bioactive metabolites in the brain, including QUIN (an NMDAR agonist) and kynurenic acid (KYNA, an α7-AChR and NMDAR antagonist). KYNA is primarily formed in astrocytes [134], which exhibit a neuroprotective function via their ability to eliminate glutamate spillover. QUIN is produced mainly by microglia [135] and has a strong excitotoxic role through enhancing NMDAR activation [136]. Microglia regulate KP balance by preferentially producing oxidative metabolites, including QUIN. Other metabolites of KP, including 3-hydroxykynurenine (3-HK) and anthralinic acid, do not directly affect neuronal activity but are involved in complicated pro-oxidation and anti-oxidation processes [137].

Recent studies have found that microglial QUIN expression was increased in the subgenual and supracallosal regions of the ACC in post-mortem brains of suicide patients with severe depression. In contrast, additional research revealed decreased QUIN in left CA1 or right CA2/3 areas of the HIP in uni- and bipolar depression patients [138]. Moreover, a meta-analysis revealed decreased KYNA and kynurenine (KYN) levels in patients with depression and increased QUIN level in antidepressant-free patients [139]. Similar results have been reported concerning the increase of QUIN in peripartum depression and adolescent MDD [140, 141]. In addition to clinical data, enhanced levels of 3-HK and QUIN have been observed in several models of depression [142, 143]. LPS and other infectious agents, including viruses, upregulate the release of inflammatory cytokines by binding to TLRs, which in turn directly and indirectly induce KP metabolism via pro-inflammatory cytokines [144]. Ketamine treatment could reduce the LPS-induced depressive-like alterations observed in the novelty suppressed feeding test (NSFT) and splash test (ST) [145]. In a genetic animal model, the level of KYNA was reduced in the PFC of Flinders sensitive line rats compared with Flinders resistant line control rats [146]. In addition, CMS contributed to QUIN production and its release from microglia in the HIP. QUIN resulted in the elevation of Glu level via NMDARs and mGluR1, as well as the increased expression of NR2B and mGluR1, which lead to depressive-like symptoms [147]. CRS exposure induced depressive-like behavior in C57BL/6 J mice, which was attributed partially to disruption of the neuroprotective/neurotoxic balance of the kynurenine metabolic pathway in the gut and brain [148]. Enzyme indoleamine 2,3-dioxygenase 1 (IDO1) is considered to be a rate-limiting enzyme for tryptophan metabolism in KP [149]. Treatment with the IDO inhibitor, 1-methyl-tryptophan, partially prevented CRS-induced depression- and anxiety-like changes [148]. The chronic forced swim test and tail suspension test in mice enhanced KYN/TRP and reduced the 5-HT/TRP ratio, which indicated activation of IDO1 [150]. In the poststroke depression (PSD) mouse model, 3-hydroxyanthranilate 3,4-dioxygenase (HAAO), QUIN, IDO1, Iba-1, and ROS were remarkably increased in the nucleus accumbens (NAc), HIP, and hypothalamus. At the same time, treatment with aripiprazole ameliorated the abnormal behaviors in PSD mice that were accompanied by decreased levels of IDO1, HAAO, QUIN, and ROS [151]. IDOInh is a specific inhibitor of IDO1 and has been shown to reverse fear learning and memory in CSDS-induced depressed mice by decreasing KYN and 3-HK levels in the blood and brain [152]. Consequently, it has been proposed that kynurenine metabolites directly interact with microglial activity, which provides a reliable target for investigating the mechanisms underlying antidepressant drugs.

Currently, numerous reports using different experimental models have indicated that antidepressants, including clinical and plant-based drugs, exert their antidepressant effects, in part, by regulating microglial phenotypes. Regulation of microglia has been proposed as a potentially effective therapeutic strategy in chronic inflammatory diseases [29, 30]. In this section, we summarized the current knowledge of antidepressants (clinical medicine and natural products) that act against microglial dysfunction in stress-induced depression models (Tables 1 and 2).

Clinical antidepressants, including selective serotonin reuptake inhibitors (SSRIs) and tricyclic antidepressants (TCAs), have exhibited effects on microglial activation and neuroinflammation [153]. Other compounds, including minocycline, melation, FCPR16, pioglitazone, iptakalim, and caffeine, also influence microglial polarization and depressive-like behavior [91], as shown in Table 1.

The SSRIs (fluoxetine, citalopram, and escitalopram) exhibit antidepressant roles in LPS, CMS, CSDS, and CWIRS models, as seen by alleviating nest-building deficits, reducing immobility time in the FST and TST, and increasing sucrose intake in the SPT (Table 1). These behavioral changes were accompanied by decreases in the microglial marker Iba-1 expression in different brain regions. In addition, pro-inflammatory factors levels (IL-1β, TNF-α, and NO) were downregulated, and anti-inflammatory factors levels (IL-4, IL-10, Arg-1, and TGF-β1) were upregulated [16, 101, 105, 154, 155]. Tricyclics, including clomipramine and imipramine, reversed depressive behaviors induced by LPS, LH, and CSDS by reducing the number of activated hippocampal microglia (Table 1). These traditional antidepressants promote the activation of anti-inflammatory microglia to release anti-inflammatory cytokines and neurotrophic factors, suggesting that reducing inflammation may be part of the function of clinical antidepressants [107, 131, 156]. Cytokines levels are strongly correlated with the efficacy of antidepressant treatment in patients. For example, higher levels of IL-6 and TNF-α are more commonly found in treatment-resistant patients than responders [157]. Nevertheless, some results revealed conflicting views that antidepressants might sometimes increase the inflammatory load in the brain. For instance, citalopram treatment elevated the levels of TNF-α and IFN-γ in the PFC, and these effects were inhibited by the anti-inflammatory agent ibuprofen [158]. Phenelzine, a monoamine oxidase inhibitor, has been shown to enhance the microglia-mediated immune responses by increasing the expressions of iNOS, TNF-α, and IL-6 in LPS-treated BV-2 cells. It has been verified that phenelzine also increased the levels of NO, TNF-α, and IL-6 via the NF-κB signaling pathway in LPS-activated primary microglia cells [159]. These differences may depend on regional brain specificity and heterogeneity among depression models and patients, but the conflicting causes and mechanisms need further study. Ketamine has exhibited potential antidepressant effects in vivo and in vitro by inhibiting microglia-mediated neuroinflammation [160]. Partial depletion of microglia with PLX3397 blocked the rapid and sustained antidepressant effects of (R)-ketamine (an isomer of ketamine), suggesting that antidepressant effects of ketamine may be partly attributed to microglial activation [161].

Minocycline is a potential agent for the treatment of depression [19]. In several studies, chronic minocycline treatment clearly alleviated depressive-like symptoms by inhibiting microglia and HPA axis hyperactivity. The powerful neuroprotective effects of minocycline were mainly mediated by modulating pro-inflammation and anti-inflammation balance in the PFC and HIP, as well as upregulating the BDNF-mediated synaptic plasticity in stress models [14, 162‐164]. Meanwhile, CMS-induced changes in behavior, hippocampal LTP, CD11b expression, NLRP3 inflammasome, BDNF, and p-GluR1 levels were restored with chronic minocycline treatment [128, 165‐168]. Peroxisome proliferator-activated receptor γ (PPARγ) is a nuclear receptor that regulates inflammation and microglial polarization and is a potential treatment target in MDD. Pioglitazone, a PPARγ agonist, significantly ameliorated depressive-like behaviors in CMS mice by regulating the expression of pro-inflammatory markers (IL-1β, IL-6, TNF-α, iNOS, and CCL2) together with anti-inflammatory markers (Ym1, Arg1, IL-4, IL-10, and TGF-β) in the HIP [169]. Other drugs that play a potential antidepressant role in modulating microglial phenotype are shown in Table 1.

Recently, the regulatory properties of some plant-derived natural compounds and formulations have shown efficacy in treating depression. In this review, the antidepressive effect of natural products such as ginsenosides (Rg1, Rg3, and Rb1), resveratrol, salvianolic acid B, and magnolol on microglia-mediated neuroinflammation are summarized in Table 2. The chemical structures of several natural compounds that may act as potential inhibitors of pro-inflammatory microglial activation in MDD are illustrated in Fig. 5.

Flavones are subgroups of flavonoids composed of a backbone of 2-phenylchromen-4-ketone, and their biological activities have been reported in vitro and in vivo [184]. Flavonoids achieve their antidepressive effects through modulating microglial activation and NLRP3 inflammasomes. Baicalin, a flavonoid found in the root of Scutellaria, could ameliorate depressive behaviors in CMS model [185, 186]. Oral administration of baicalin (20, 40 mg/kg/day) for 3 weeks significantly improved sucrose intake, locomotor activities, and behavioral despair in CMS rats. The mechanism of baicalin against CMS-induced depression was partially achieved by increasing levels of DCX, neuron-specific enolase (NSE), and BDNF, reducing oxidative stress, and modulating the GSK3/NF-κB/NLRP3 signal cascade in the HIP [185]. In addition, baicalin supplementation (30, 60 mg/kg/day, p.o.) for 3 weeks also inhibited TLR4 activation and pro-inflammatory cytokine secretion in CMS mice. Furthermore, it was confirmed that the inhibition might be realized through PI3K/AKT/FoxO1 pathway in LPS-stimulated BV2 cells [186]. Quercetin is a flavanol present in apples, onions, and berries [37]. Treatment with quercetin (40, 80 mg/kg/d, p.o.) for 2 weeks significantly reversed OBX-induced increase in immobility time in the FST, possibly by inhibiting TNF-α, IL-6, and caspase-3 in the HIP and cortex. Co-treatment with minocycline (25 mg/kg, p.o.) significantly potentiated its protective effects when compared to treatment with quercetin alone [187]. Treatment with quercetin (40 mg/kg/day, p.o.) for 2 weeks alleviated depressive-like behavior in the SPT and FST of LPS-challenged rats via regulation of BDNF-related imbalance of Copine 6 and TREM1/2 in the HIP and PFC [188].

Terpenoids, especially triterpenoids, monoterpenes and sesquiterpenes, are the most abundant natural compounds found in several vegetables and fruits. Recently, some terpenoids have been reported to have preventive efficacy against neurological diseases by modulating the activation of microglia, especially on depression. Saponins, which are triterpene glycosides, are abundant in some plants, and their general biological activities have been summarized [189]. Saikosaponin-D is a triterpenoid saponin derived from Radix bupleuri and exhibits anti-inflammatory, anti-oxidative, and other pharmacological activities [190]. Previous studies have suggested that treatment with saikosaponin-D (0.75, 1.50 mg/kg/day, p.o.) for 3 weeks remarkably ameliorated the CMS-induced depressive-like behaviors in rats, mainly by improving HPA axis functions and hippocampal neurogenesis [191]. Moreover, additional research confirmed that saikosaponin-D administration ameliorated LPS-induced depressive-like behaviors, as shown by increased sucrose consumption in the SPT and decreased immobility time in the TST and FST. These performances appear to be mediated by inhibition of microglial activation and regulation of the high mobility group box 1 (HMGB1)/TLR4/NF-κB signaling pathway [190]. Gypenosides, the major ingredients of Gynostemma pentaphyllum, exert a neuroprotective function in the CNS. Treatment with gypenosides (50, 100 mg/kg/day, p.o.) for 4 weeks significantly relieved depressive-like behaviors of CMS mice in the TST and FST, which was mediated, in part, by inhibiting microglial activation and NF-κB signaling and increasing BDNF levels in the HIP [192, 193].

Ginsenosides are one of ginseng’s most biologically active ingredients and have a triterpenoid glycoside structure [194]. Direct and indirect evidence support that some ginsenosides (such as ginsenoside Rg3, Rb1, and Rg1) induce the anti-inflammatory actions of macrophages and microglia [195]. Ginsenoside Rb1 is a typical 20(S)-protopanaxadiol-type saponin and exerts significant antidepressive effects in chronic stress models [73, 196, 197]. It is known that Rb1 regulates microglial activation, protects neurons from inflammatory and oxidative damage, and promotes neurogenesis [198, 199]. Recently, treatment with Rb1 (20 mg/kg/d, p.o.) in CMS-treated mice for consecutive 4 weeks alleviated depressive-like behaviors in the SPT, TST, and FST mainly through PPARγ-mediated transitions in microglial phenotype (TNF-α, IL-1β, TGF-β, and Arg-1) in the HIP and cortex [73]. Rb1 attenuated a decrease in BDNF and the ratio of p-AKT/AKT expression, and increased IL-1β, TNF-α, and Iba-1 levels in the HIP in a CRS model. In line with in vivo reports, Rb1 lowered the protein expressions of IL-1β and TNF-α in BV-2 cells [197]. In addition to Rb1, ginsenoside Rg1 and Rg3 exhibited potential anti-inflammatory effects, which protected against microglial activation in stress-injured rodents [200, 201]. A previous study reported that chronic pretreatment with ginsenoside Rg1 (40 mg/kg/d, i.p.) for 5 weeks significantly suppressed inflammatory responses by alleviating microglial and astrocyte activation through decreased overexpression of IL-1β, IFN-γ, and TNF-α. These effects were accompanied by attenuation of dendritic spine and synaptic defects, and upregulation of synaptic-related proteins in the ventral medial prefrontal cortex (vmPFC). Ginsenoside Rg1 inhibited neuronal apoptosis by increasing Bcl-2 and decreasing cleaved caspase-3 and caspase-9 expressions after CMS exposure. Furthermore, ginsenoside Rg1 increased the nuclear factor erythroid 2-related factor (Nrf2) and inhibited p38 mitogen-activated protein kinase (p-p38 MAPK) and p65-NF-κB activation in the vmPFC [200]. Moreover, ginsenoside Rg3 at oral doses of 20 and 30 mg/kg attenuated upregulation of TNF-α, IL-1β, and IL-6 mRNA expression in brains of C57BL/6 mice after systemic LPS injection. Morphological activation of microglia, and Iba-1, cyclooxygenase-2 (COX-2), iNOS protein expression was reduced with Rg3 treatment [201]. Our previous studies also found that ginsenosides Rd and Re (10, 20, or 40 mg/kg, p.o.) have potentially neuroprotective and anti-inflammatory properties, as manifested by significantly reducing the expression of hippocampal pro-inflammatory factors and NLRP3 inflammasome related protein, as well as enhancing endogenous antioxidant factor Nrf2, and mediating PI3K-AKT and BDNF signaling pathways in mice exposed to CRS [129, 202]. Taken together, the anti-inflammatory effects of ginsenosides have been confirmed, and the negative regulation of pro-inflammatory cytokines and enzymes has been found to underlie the anti-inflammatory properties of ginsenosides in pro-inflammatory microglia in depression. However, additional studies are needed to investigate whether other ginsenosides or their metabolites can relieve depressive symptoms through modulating microglial phenotype and the molecular mechanisms involved in neuroinflammation.

Catalpol, an iridoid glucoside, is primarily isolated from Radix rehmannia and is commonly used as a traditional Chinese medicine [203]. Previous studies have been reported that catalpol exhibits a wide range of pharmacological effects, including exhibiting anti-diabetic, anti-tumor, anti-inflammatory, and antioxidant activities. Growing evidence indicates that catalpol has a robust antidepressant effect that acts through its anti-inflammatory and anti-oxidative properties in vitro and in vivo [203, 204]. Catalpol (5, 10, or 20 mg/kg/day, i.g.) administration for 5 weeks ameliorated CMS-induced depressive-like behavior in the SPT, and its underlying mechanisms might be at least partially ascribed to reducing HPA axis dysfunction, upregulating BDNF and its specific binding receptor tyrosine kinase B (TrkB), downregulating COX-2 expression, thus reducing prostaglandin E2 level in the brain [204]. It was also confirmed that catalpol at 20 mg/kg decreased the expression of NLRP3 inflammasome-associated proteins and inhibited pro-inflammatory microglial polarization in the HIP (IL-1β, TNF-α, and iNOS) [205]. Geniposide, a type of iridoid glycoside extracted from the ripe fruit of Gardenia jasminoides Ellis, exhibits numerous bioactivities, including anti-diabetic, anti-oxidative, and anti-inflammatory actions [206]. It has been demonstrated that administration of geniposide (10, 40 mg/kg/day, p.o.) for 1 or 3 weeks ameliorated LPS- and CMS-induced depressive-like behavior in the SPT, TST, and FST by regulating the Bruton’s tyrosine kinase (BTK)/TLR4/NF-κB, BDNF/TrkB, and BTK/JAK2/STAT1 signaling pathways [207, 208]. Also, the beneficial effect of geniposide on anxiety- and depressive-like behaviors in mice may possibly act through induction of microglial polarization towards anti-inflammatory phenotype and inhibition of IL-6 and TNF-α release [208].

Arctigenin and its glycoside, arctiin, are the major active ingredients of the dried ripe fruit of Arctiumlappa L. Arctiin can be metabolized into arctigenin by human intestinal microflora after oral consumption [209]. Arctiin has antidepressive effects that appear to act through inhibition of the NF-κB mediated HMGB1/TLR4 and TNF-α/TNF receptor 1 (TNFR1) pathways, which consequently attenuated microglial activation and neuroinflammation [210]. Similarly, arctigenin showed antidepressive effects by inhibiting microglial activation and neuroinflammation via HMGB1/TLR4/NF-κB and TNF-α/TNFR1/NF-κB signaling pathways [171]. Salvianolic acid B (SalB) is one of the phenolic acid compounds derived from Salvia miltiorrhiza, which has various pharmacological effects, including anti-inflammatory, antioxidant, and anti-apoptotic [211, 212]. CMS mice treated with SalB (20 mg/kg/day, i.p.) for 3 weeks significantly reversed decreased sucrose preference index in the SPT and increased immobility time in the FST and TST. In addition, the decreased expression of IL-1β and TNF-α and increased expression of IL-10 and TGF-β were accompanied by increased apoptosis (cleaved-caspase 3⁺) and microglial activation (Iba-1⁺) in the HIP and cortex that were reversed after SalB treatment [213]. Curcumin is a representative natural phenolic compound of turmeric and a potential candidate for regulation of brain function due to its antioxidant and anti-inflammatory properties. In a CRS model, systemic administration of curcumin (10, 20, or 30 mg/kg, i.p.) daily for 3 weeks significantly attenuated oxidative stress and lipid peroxidation, prevented apoptosis, and increased antioxidant defense activity [214]. Curcumin (10 mg/kg, i.p.) also was beneficial for maintaining improved memory and mood functions after 30 days of daily therapy in a Gulf War Illness (GWI) with stress-treatment model. At the molecular level, enhanced numbers of DCX⁺ cells, decreased numbers of Iba-1/ED-1⁺ cells, and elevated antioxidant genes with normalized mitochondrial respiration might be the basis of the regulatory mechanism that is mediated by curcumin treatment [215]. Magnolol, a hydroxylated biphenyl compound extracted from Magnolia tree, has endothelial cell protective, antioxidant, and anti-inflammatory functions [216]. It is noteworthy that magnolol exhibits antidepressant effects, as evident by increased sucrose consumption and decreased the immobility time in the SPT, TST, and FST. In an OBX model, magnolol (50, 100 mg/kg/day, p.o.) treatment for 2 weeks produced antidepressant-like effects by enhancing neurogenesis (as indicated by BrdU positive cells) in the HIP [217]. In a CMS model, treatment with magnolol (20, 40 mg/kg/day, p.o.) for 4 weeks increased BDNF expression and normalized serotonergic system [218]. In agreement with previous studies, the protective effect of magnolol (50, 100 mg/kg/day, p.o.) given for 3 weeks to CMS-treated female mice also appeared to be associated with inhibition of pro-inflammatory microglia (TNF-α, IL-1β, and IL-6) and activation of anti-inflammatory microglia (IL-4, IL-10, Arg1, Ym1, Fizz1, and Klf4) via Nrf2/hemeoxygenase (HO-1)/NLRP3 signaling pathway [219].

Resveratrol is one of the most well-known dietary stilbenoids. Resveratrol confers health benefits due to its antioxidant, anti-inflammatory, anti-aging, and immune-regulatory activities [220]. Resveratrol is found in the skin of red grapes, peanuts, and other medicinal plants. Recent studies have found that resveratrol ameliorated depressive-like behaviors in several animal models [221‐223]. In a LPS model, mice were treated for 2 weeks with resveratrol (20 mg/kg/day, i.p.), which abrogated the increased immobility in the FST and TST. Immunohistochemical staining revealed that resveratrol reversed increased microglial activation (Iba-1⁺) and inhibition of neurogenesis (BrdU/DCX⁺) in the DG [224]. Moreover, 3 doses of resveratrol (20 mg/kg/day, i.p.) promoted the activation of sirtuin type 1 and blocked the decline of hippocampal neurogenesis triggered by ethanol exposure during early postnatal life [225]. The carotenoids crocin was able to regulate microglial activation associated with neurological disorders. Crocin decreased the expression of LPS-induced NO, TNF-α, and ROS production in BV-2 cells and improved locomotor activity, sucrose intake, and reduced immobility time in LPS-treated Kunming mice [226]. The health benefits associated with crocin might be due to its potent ability to regulate the NLRP3 inflammasome and NF-κB, as well as anti-inflammatory phenotypic conversion [226]. Moreover, emerging data have shown that plant-derived compounds can exert neuroprotective effects that prevent neurological disorders in depression. For instance, ganoderic acid A [227], asperosaponin VI [228], 20(S)-protopanaxadiol [229], scutellarin [230], astragalin [231], hesperidin [232], 5-O-methylvisammioside [233], ginkgolide B [234], leonurine [235], ferulic acid [236], and (+)-sesamin [69] were found to inactivate microglia in stress-induced depression models. Recent information concerning the antidepressant activity of phytochemicals through regulation of microglial polarization is illustrated in Table 2. In addition, several herbal extracts and traditional formulations reported to improve depression-like behavior by regulating microglia are shown in Table 2 [237‐245].