Abstract

Depression is one of the leading causes of disability worldwide, yet its underlying mechanisms remain elusive. Both neuroinflammation and ferroptosis are known contributors to the disease, and understanding their interplay is critical for unraveling the pathogenesis of depression. This review explores the relationship between neuroinflammation and ferroptosis in the context of depression. We first delineate the core mechanisms of each process. Subsequently, we focus on the key signaling pathways bridging these processes, including cyclic GMP-AMP synthase (cGAS) stimulator of interferon genes (STING), nuclear factor kappa-B (NF-κB), Janus kinase (JAK)/signal transducer and activator of transcription (STAT) and nuclear factor erythroid 2-related factor 2 (Nrf2), and elucidate how this inflammation-ferroptosis vicious cycle contributes to depression. Finally, we highlight therapeutic agents targeting both processes, suggesting novel treatment directions. By showing that inflammation and ferroptosis form a vicious cycle, this work offers a clearer perspective on the pathogenesis of depression and identifies specific therapeutic targets for breaking this pathological loop.

1 Introduction

Depression is a common mental disorder, with a global prevalence that has increased significantly over the past three decades, especially in Asia, North Africa, and the Middle East (Liu et al., 2020). Depressive disorders rank as the third most prevalent mental disorders worldwide (GBD 2019 Mental Disorders Collaborators, 2022). However, the clinical utility of conventional antidepressants is often limited by modest efficacy and a range of adverse effects (Cowen, 2008; Pigott et al., 2010; Strawn et al., 2023; Zisook et al., 2006). For instance, selective serotonin reuptake inhibitors (SSRIs), the most commonly prescribed antidepressants, carry a black-box warning for increased suicidal risk, and frequently cause sexual dysfunction in up to 80% of female patients (Fornaro et al., 2019; de Aquino et al., 2025). Besides, most antidepressants share common challenges, including a marked therapeutic delay of several weeks, limited bioavailability, poor blood–brain barrier permeability, and instability in the gastrointestinal tract (Ferrari and Villa, 2017; Văruț et al., 2025). Moreover, approximately one-third of patients exhibit treatment-resistant depression (TRD) and fail to respond to existing pharmacotherapies, while some patients, particularly children and adolescents, may experience spontaneous improvement even without treatment (Richardson et al., 2025; Kennard et al., 2009). These observations collectively suggest that the underlying pathophysiology of depression needs to be further investigated. Thus, elucidating the mechanisms of depression and developing novel therapeutics with distinct mechanisms of action remain a critical priority.

While the exact etiology of depression is still under investigation, inflammation has emerged as a significant factor in its development (Barton, 2008). Stress can induce inflammation, which in turn promotes depressive-like behaviors in some patients (Bierhaus et al., 2003; Irwin and Cole, 2011; Kiecolt-Glaser et al., 2015; Shelton and Miller, 2010). Furthermore, anti-inflammatory treatments can alleviate depressive-like behaviors, indicating the potential antidepressant effects of anti-inflammatory drugs (Kang et al., 2011; Liu H. et al., 2023; Guo et al., 2019). Meanwhile, ferroptosis—an iron-dependent programmed cell death (PCD) first reported by Dixon et al.—has emerged as a key player in depression, with studies documenting its presence in neurons and glia and showing that suppressing it alleviates depressive-like behaviors (Dang et al., 2022; Shen et al., 2024b; Mao et al., 2024; Li E. et al., 2023). Emerging evidence suggests that these two pathological processes are not independent; rather, they engage in a self-amplifying vicious cycle—inflammation potentiates ferroptosis, and ferroptotic products and intracellular iron overload in turn fuel neuroinflammation (Li E. et al., 2023; Xin et al., 2020; Yu et al., 2020; Wang et al., 2019; Li et al., 2019). Elucidating this vicious cycle is fundamental to understanding depression pathogenesis and to identifying new therapeutic targets.

This review begins by delineating the core mechanisms of neuroinflammation and ferroptosis individually and then proceeds to explore their molecular interplay within the central nervous system (CNS). By framing these processes as a vicious cycle, we aimed to provide an integrated perspective on how neuroinflammation and ferroptosis collectively drive depressive pathology. Finally, we reported therapeutic agents that target this vicious cycle, thereby providing novel therapeutic strategies for depression.

2 Neuroinflammation in depressive disorder: multidimensional pathogenic mechanisms

Neuroinflammation is an evolutionarily conserved defense mechanism in the CNS, triggered by insults such as infection or injury and mediated through the coordinated activation of resident glial cells and recruited immune cells (Kopp et al., 2022). These cells release proinflammatory mediators including cytokines, chemokines, and reactive oxygen species to clear threats and restore tissue homeostasis. While acute neuroinflammation is essential for CNS protection, its persistent dysregulation transforms into a pathogenic force that drives neuronal damage and underpins a broad spectrum of neurological and psychiatric conditions. Critically, this maladaptive neuroinflammatory state has been increasingly implicated in the pathophysiology of major depressive disorder (MDD) (Tastan and Heneka, 2024; Sălcudean et al., 2025). This link is thought to arise from prolonged immune activation, which disrupts several brain processes central to mood regulation. This activation is characterized by the sustained activation of neuroglial cells, such as microglia and astrocytes (Yang et al., 2020). Therefore, the dynamic phenotypic shift of glial cells, particularly the interplay between microglia and astrocytes, is emphasized as a central mechanism of neuroinflammation (Figure 1).

Interactive network of glial and infiltrated immune cells in central neuroinflammation in depression. This schematic illustrates the central cellular interactions that drive sustained neuroinflammation within the central nervous system (CNS) in depression. At the core of this network is a self-perpetuating, pathological cycle: Activated “M1 microglia” release pro-inflammatory cytokines such as tumor necrosis factor-alpha (TNF-α). These cytokines push astrocytes toward a neurotoxic “A1 astrocyte” phenotype. In turn, A1 astrocytes secrete additional inflammatory mediators, including chemokines like CCL2 and CXCL10, which further activate microglia, closing a positive feedback loop. This cycle is amplified by the recruitment of “peripheral immune cells”—for example, IL-17-producing T cells—across the blood–brain barrier. “Pericytes,” which surround blood vessels in the CNS, also adopt a pro-inflammatory profile, releasing TNF-α and CCL2, thereby enhancing both immune cell infiltration and glial activation. Opposing this pro-inflammatory cascade is a resolution pathway driven by mediators such as IL-4, IL-10, and transforming growth factor-beta (TGF-β). These signals promote protective “M2 microglia” and “A2 astrocyte” phenotypes, which support tissue repair and suppress inflammation. However, in major depressive disorder, this reparative axis is often overwhelmed. The persistent dominance of M1/A1 activation and peripheral immune infiltration leads to sustained neuroinflammation, which is thought to contribute to maladaptive behavioral responses, including anhedonia, psychomotor retardation, and mood dysregulation—core features of depression. Home for Researchers. (2022). FigDraw(Version 2.0) [Computer software]. https://www.figdraw.com/

Under physiological conditions, microglia maintain central nervous system (CNS) homeostasis through a dynamic and context-dependent spectrum of functional states, traditionally described as a balance between pro-inflammatory and protective responses (Tastan and Heneka, 2024). In neuropsychiatric disorders as depression, this balance persistently shifts toward a pro-inflammatory phenotype. Damage-associated molecular patterns, such as high mobility group box 1 (HMGB1), initiate this phenotypic shift by engaging signaling pathways like toll-like receptor 4 (TLR4) and the receptor for advanced glycation end products (RAGE), which in turn activate downstream effectors, including nuclear factor kappa-B (NF-κB) and the NOD-like receptor thermal protein domain associated protein 3 (NLRP3) inflammasome (Tastan and Heneka, 2024; Afridi and Suk, 2023; Ajoolabady et al., 2025). Concurrently, impaired function of regulatory receptors, such as triggering receptor expressed on myeloid cells 2 (TREM2), reduces phagocytic clearance and attenuates anti-inflammatory signaling, thereby impairing endogenous resolution pathways (Ajoolabady et al., 2025). Upon activation, microglia drive neuroinflammation through the release of mediators, including tumor necrosis factor-alpha (TNF-α), interleukin-1 alpha (IL-1α), and complement component 1q (C1q), which promote the conversion of astrocytes into a neurotoxic A1-reactive phenotype (Tastan and Heneka, 2024). Together with pericytes and infiltrating peripheral immune cells, these cells sustain a chronic pro-inflammatory environment characterized by elevated cytokines (e.g., interleukin-1 beta [IL-1β], TNF-α, interleukin-6 [IL-6]), complement system activation (e.g., C1q, complement [C3]), and persistent inflammasome activity. These cascades collectively drive synaptic loss, axonal damage, dysregulated calcium signaling, impaired long-term potentiation, and altered neural network connectivity (Ajoolabady et al., 2025). While useful as a simplified framework that characterizes microglia as either pro-inflammatory (M1) or protective/repair-oriented (M2), the classical M1/M2 dichotomy does not fully capture the diverse and dynamic activation states of microglia, which often exist along a phenotypic continuum between pro-inflammatory and protective functions (Tastan and Heneka, 2024; Yang et al., 2020). The resulting structural and functional impairments in neuronal integrity contribute to cognitive deficits, mood dysregulation, and other neuropsychiatric symptoms, thereby reinforcing a self-perpetuating cycle of neuroimmune dysregulation.

Astrocytes are now recognized as active regulators of central nervous system (CNS) homeostasis and neuroinflammation (Giovannoni and Quintana, 2020). Under physiological conditions, they contribute to blood–brain barrier (BBB) integrity, ion and fluid balance through aquaporin-4 (AQP-4) channels, synaptic modulation via glutamate transporters (excitatory amino acid transporters [EAATs], including glutamate transporter-1 [GLT-1] and glutamate–aspartate transporter [GLAST]), and metabolic support through lactate shuttling (Yang et al., 2020). In neuroinflammatory contexts, astrocytes undergo reactive transformation, acquiring phenotypic states that may exert either detrimental or protective effects, depending on microenvironmental signals. Microglia play a crucial role in orchestrating this phenotypic shift: upon activation, they release inflammatory mediators such as IL-1α, TNF-α and C1q, which induce a pro-inflammatory reactive astrocyte phenotype (often termed “A1-like” or neurotoxic reactive state) (Zhao et al., 2024; Lee et al., 2023; Jha et al., 2019). Such reactive astrocytes are characterized by reduced glutamate uptake capacity and increased release of pro-inflammatory cytokines, including TNF-α and IL-1β, which contribute to synaptic dysfunction and neuronal injury. In contrast, under reparative conditions, microglia may support an alternative reactive astrocyte phenotype (“A2-like” or neuroprotective state), associated with the secretion of anti-inflammatory mediators such as interleukin-10 (IL-10), interleukin-4 (IL-4), transforming growth factor-β (TGF-β), and neurotrophic factors (Zhao et al., 2024). This phenotypic transition reflects a bidirectional crosstalk: activated astrocytes, in turn, modulate microglial activity through chemokines such as monocyte chemoattractant protein-1 (MCP-1/CCL2) and C-X-C motif chemokine ligand 10 (CXCL10), thereby amplifying or attenuating neuroinflammatory cascades (Yang et al., 2020). Thus, the dynamic interaction between microglia and astrocytes, mediated by specific cytokine and chemokine signaling pathways, critically determines neuroinflammatory trajectories and influences both degenerative and regenerative processes in CNS disorders (Jha et al., 2019; Liu et al., 2020).

Moreover, pericytes and infiltrating peripheral immune cells also contribute to the neuroinflammatory cascade. They influence the initiation, progression, and resolution of neuroinflammation through multiple mechanisms, including the release of inflammatory mediators, regulation of immune cell infiltration, interactions with microglia and endothelial cells, inflammasome activation, extracellular matrix degradation, and phenotypic plasticity (Tastan and Heneka, 2024; Liu et al., 2020). In summary, for patients with depression, neuroinflammation—primarily driven by the chronic activation of glial cells—induces a cascade of structural and functional alterations across key emotion- and cognition-regulating brain regions (Sălcudean et al., 2025; Troubat et al., 2021). For example, it leads to reduced hippocampal volume and impaired neurogenesis; prefrontal cortical gray matter atrophy and hypoconnectivity; hyperactivity of the amygdala and heightened connectivity; structural alterations and dysfunctional conflict processing in the anterior cingulate cortex; and dysregulated dopaminergic signaling and blunted reward response in the nucleus accumbens (Sălcudean et al., 2025). These coordinated damages across multiple brain regions collectively disrupt the neural circuits governing emotion regulation, cognitive execution, and motivational reward, ultimately resulting in the core clinical symptoms of persistent low mood, anhedonia, executive dysfunction, and cognitive impairment in patients (Troubat et al., 2021; Reyes-Martínez et al., 2023). Importantly, these neuroinflammatory damages may be further aggravated by a vicious cycle with ferroptosis, as discussed in the later sections.

In depression, neuroinflammation is induced by psychosocial stress through various mechanisms including activation of microglia, activation of the sympathetic nervous system (SNS), and glucocorticoid resistance (Miller and Raison, 2016; Amit et al., 2009; Berkenbosch et al., 1987; Miller et al., 2002). The activation of microglia directly results in the secretion of proinflammatory cytokines in the CNS, which may attract peripheral monocytes to further promote neuroinflammation (Miller and Raison, 2016; Tang and Le, 2016). On the other hand, activation of the SNS in turn activates peripheral leukocytes to promote expression of NF-κB-mediated pro-inflammatory immune response genes such as IL-1β, IL-6, and TNF-α via adrenaline and noradrenaline, which may generate an inflammatory environment in the periphery (Amit et al., 2009; Cole et al., 2010; Grebe et al., 2010). Furthermore, glucocorticoid resistance reduces the anti-inflammatory effects of glucocorticoids, which aggravates inflammation (Berkenbosch et al., 1987; Miller et al., 2002; Sapolsky et al., 1987; Besedovsky et al., 1986; Duman and Aghajanian, 2012). Finally, the peripheral inflammatory cytokines can not only cross the BBB but also bind to receptors on endothelial cells to activate the production of prostaglandins (PGs), which promotes neuroinflammation (Schiepers et al., 2005; Friedman, 2001; Utsuyama and Hirokawa, 2002; Watkins et al., 1995). Thus, neuroinflammation can be established under chronic social-environmental threats.

Inflammatory cytokines affect neurotransmission through several pathways, a process which is related to neuronal dysfunction and synaptic plasticity in depression. For example, inflammatory cytokines can induce the oxidative loss and decrease the concentration of tetrahydrobiopterin (BH4) (Felger et al., 2013; Werner et al., 2003). As an enzymatic cofactor in the synthesis of monoamine neurotransmitters, including serotonin and dopamine, reduced BH4 levels contribute to decreased 5-Hydroxytryptamine (5-HT) and dopamine (DA) levels, which induces depressive-like behaviors (Nasser et al., 2014; Kwak et al., 2013; Zeng et al., 2004; Homma et al., 2011). Furthermore, inflammatory cytokines induce the activation of indoleamine 2,3-dioxygenase (IDO), an enzyme that catalyzes the production of kynurenine from tryptophan, thereby reducing the alternative metabolic pathway that yields 5-HT (Raison et al., 2010; Maes et al., 2011). Furthermore, activation of the kynurenine pathway promotes the production of quinolinic acid, an excitotoxic compound that promotes glutamate release and blocks glutamate reuptake (Steiner et al., 2011; Tavares et al., 2002). Besides, inflammatory cytokines also directly affect extracellular levels of glutamate by downregulating the expression of glutamate reuptake pump on astrocytes and promoting glutamate release of astrocytes (Tilleux and Hermans, 2007). Excessive glutamate over activates the postsynaptic glutamate receptors, disrupting intracellular calcium homeostasis and contributing to in neuronal death (Simões et al., 2018; Heath and Shaw, 2002; Scheefhals and MacGillavry, 2018). Furthermore, overactivation of N-Methyl-D-aspartic acid (NMDA) receptors by glutamate reduces brain-derived neurotrophic factor (BDNF) production, resulting in decreased neurogenesis. Therefore, neuroinflammation induced in depression disrupts neurotransmission and thus promotes the development of depression (Hardingham et al., 2002; Koo et al., 2010; Goshen et al., 2008).

Neuroinflammation plays a critical role in hippocampus atrophy as well (Zhang et al., 2024; Bajrami et al., 2022; Cherbuin et al., 2019; Semmler et al., 2013). On the one hand, inflammation reduces adult hippocampal neurogenesis (Borsini et al., 2021; Yirmiya and Goshen, 2011; Chen P. et al., 2023). On the other hand, inflammation induces PCD of various cells in the hippocampus and prefrontal cortex (PFC), including neuron, glial cells, and neural stem cells (NSCs) (Li W. et al., 2023; Li et al., 2026; Nikolopoulos et al., 2023). As the hippocampus and adult hippocampal neurogenesis (AHN)—a neuronal maturation process dependent on NSCs—are important for learning, cognition, and memory, disrupted hippocampal function and suppressed neurogenesis contribute to the development of depression (Yirmiya and Goshen, 2011; Hitti and Siegelbaum, 2014; Time, 2003; Tartt et al., 2022; Ballesteros et al., 2025). Notably, hippocampal atrophy and impaired neurogenesis are not only consequences of neuroinflammation but also sites where ferroptosis can be triggered, potentially establishing a local vicious cycle that exacerbates structural and functional deficits.

3 Ferroptosis: a novel player in depression pathogenesis

Dixon and his team introduced the concept of ferroptosis in 2012, characterizing it as an iron-dependent mode of programmed cell death marked by the build-up of lipid hydroperoxides (Dixon et al., 2012). The morphological features of ferroptosis primarily involve mitochondrial alterations including rupture of the mitochondrial membrane, reduced mitochondrial size with increased mitochondrial membrane densities, and loss of mitochondrial cristae (Li J. et al., 2020). In addition, disruption of plasma membrane integrity occurs in cells undergoing ferroptosis, as evidenced by propidium iodide (PI) staining (Liu et al., 2020). Nevertheless, the nucleus of ferroptotic cells maintains basic structural integrity (GBD 2019 Mental Disorders Collaborators, 2022).

Iron homeostasis and its dysregulation are the core of ferroptosis. First, transmembrane iron transportation is a key regulator of intracellular iron homeostasis. Fe3+ is bond by transferrin (TF) in the plasma and combines with transferrin receptor protein 1 (TFR1) on the cell membrane, facilitating cellular entry via endocytosis. Protons are pumped into the endosome, lowering the pH and triggering the release of Fe3+ from the TF-TFR1 complex. The released Fe3+ is then reduced to Fe2+ by six-transmembrane epithelial antigen of prostate 3 (STEAP3) and exported from the endosome into the cytoplasm by divalent metal transporter1 (DMT1). Following iron release, the TF-TFR1 complex is recycled back to the cell membrane via the endosome, which recycles TF and TFR1 (Candelaria et al., 2021). Second, cytoplasmic Fe2+ can be exported from cells by ferroportin (FPN), a process inhibited by hepcidin, which blocks and ubiquitinates FPN. Intracellular Fe2+ serves dual roles: it can be securely stored via ferritin binding in the cytoplasm or mitochondria, or it can function as a cofactor in the synthesis of heme and iron–sulfur clusters (Kowdley et al., 2021; Richardson et al., 2010). However, these intracellular iron storage mechanisms are reversible. For example, cytosolic ferritin can release Fe2+ through a process mediated by nuclear receptor coactivator 4 (NCOA4)-induced autophagy, known as ferritinophagy (Gao et al., 2016). Moreover, although the autophagy of mitochondria (mitophagy) may sequester iron in mitophagosomes to prevent iron from releasing into the cytoplasm in the early stages of iron overload, extensive mitophagy can conversely trigger Fe2+ liberation, elevate cytosolic Fe2+ levels, and amplify ferroptosis (Yu et al., 2022; Rademaker et al., 2022).

In the presence of Fe2+, ferroptosis is primarily induced through two pathways: the Fenton reaction-dependent pathway and the Fe2+-dependent enzymatic pathway (Murphy, 2009; Yang et al., 2016). Both pathways utilize polyunsaturated fatty acid (PUFA)-containing phospholipids (PLs) as substrates. PUFA-PLs are synthesized from PUFA through the sequential action of acyl-coenzyme A (CoA) synthetase long-chain family member 4 (ACSL4) and lysophosphatidylcholine acyltransferase3 (LPCAT3). These PUFA-PLs can then react with hydroxyl radicals generated by the Fenton reaction (Fe2+ reacting with hydrogen peroxide [H2O2]; H2O2 is primarily derived from the superoxide leaking from mitochondrial electron transport chain [ETC] complex I and III) to form PUFA radicals (PUFA·). PUFA· subsequently reacts with oxygen to yield PUFA peroxyl radicals (PUFA-OO·) and, ultimately, phospholipid hydroperoxides (PL-OOH) (Yang et al., 2016). Alternatively, PUFA-PLs can directly react with oxygen to yield PL-OOH in a process catalyzed by Fe2+ and lipid-peroxidizing enzymes such as lipoxygenases [ALOXs] and cytochrome P450 oxidoreductase [POR] (Yang et al., 2016; Koppula et al., 2021).

Glutathione peroxidase 4 (GPX4) is an essential peroxidase that suppresses ferroptosis, requiring glutathione (GSH) as a cofactor to catalyze the reduction of PL-OOH. GSH synthesis requires cystine, which is imported by cystine/glutamate antiporter (system XC−), a system composed of solute carrier family 7 member 11 (SLC7A11) and solute carrier family 3 member 2 (SLC3A2)(Dixon et al., 2012). Another system that suppresses ferroptosis, FSP1-CoQ10-NAD(P)H, operates in parallel to the GSH/GPX4 pathway. Ferroptosis suppressor protein 1 (FSP1) reduces coenzyme Q10 (CoQ) by utilizing nicotinamide adenine dinucleotide phosphate hydrogen (NADPH) to generate CoQH2, a radical-trapping antioxidant (RTA) that traps lipid peroxyl radicals and thus inhibits ferroptosis (Doll et al., 2019; Bersuker et al., 2019).

Emerging evidence implicates ferroptosis in the pathogenesis of depression (Dang et al., 2022; Shen et al., 2024b). For instance, hippocampal ferroptosis and depressive-like behaviors were observed in lipopolysaccharides (LPS)-induced mice, whereas treatment with ferrostatin-1 (Fer-1), a ferroptosis inhibitor, alleviated depressive-like behaviors of chronic unpredictable mild stress model (CUMS)-induced mice (Mao et al., 2024; Li E. et al., 2023). In addition, ferroptosis contributes to dysfunction of various CNS cell types in depression. Therefore, ferroptosis and neuroinflammation are not parallel pathological features; they are mechanistically coupled, forming a bidirectional vicious cycle that we will dissect in the following sections.

4 Bidirectional crosstalk between neuroinflammation and ferroptosis in depression

Although the individual roles of neuroinflammation and ferroptosis in the development of depression have received considerable attention, the bidirectional crosstalk between these processes and their specific molecular pathways remains relatively unexplored. By understanding these connections, the complex mechanism of depression pathogenesis can be further elucidated, thereby facilitating the development of novel antidepressant therapies that target both processes. The crosstalk between inflammation and ferroptosis has been systematically reviewed, however, its specific role in the CNS, especially in depression, has not been thoroughly addressed. This section focuses on the molecular wiring that sustains the vicious cycle between neuroinflammation and ferroptosis. By mapping how inflammatory pathways promote ferroptosis and how ferroptotic signals re-ignite inflammation, we can identify the leverage points to break the cycle (Figure 2).

Main pathways linking ferroptosis and inflammation in the central nervous system. (A) The cyclic GMP-AMP synthase (cGAS)/stimulator of interferon genes (STING) pathway is activated by damage-associated molecular patterns (DAMPs) such as double-stranded DNA (dsDNA). STING activates nuclear factor kappa-B (NF-κB), downregulates dihydroorotate dehydrogenase (DHODH), glutathione peroxidase 4 (GPX4), and solute carrier family 7 member 11 (SLC7A11). Activation of NF-κB promoting further production of proinflammatory cytokines and suppression of anti-ferroptotic genes. DHODH is a mitochondrial enzyme that suppresses lipid peroxidation and ferroptosis. GPX4 catalyzes the reduction of PL-OOH, which needs cystine/glutamate antiporter (system XC−) (a system composed of SLC7A11 and solute carrier family 3 member 2) to transport the ingredients for its cofactor. STING thus promotes neuronal ferroptosis via inhibiting DHODH and GPX4 system. (B) Inflammatory cytokines combine to TLR to activate IKK and in turn NF-κB, which not only promotes inflammation but also downregulates GPX4. (C) Nuclear factor erythroid 2-related factor 2 (Nrf2)/ heme oxygenase-1 (HO-1) axis not only inhibits NF-κB to supress inflammation but also upregulates GPX4 and SLC7A11 to reduce phospholipid hydroperoxides (PL-OOH) and prevent ferroptosis. However, this axis was reported to be inhibited by HMGB1-Serotonin receptor 7 (5HT7R) axis, resulting ferroptosis of M2 microglia. In contrast, edaravone (EDV) activates Nrf2/HO-1 axis via SIRT1, which alleviates neuronal ferroptosis. Besides, HMGB1 activates the upregulation of hepcidin via various pathways. Since hepcidin inhibits Fe2+ export, HMGB1 is a critical target for intracellular iron accumulation. (D) The Janus kinase (JAK)/signal transducer and activator of transcription (STAT) is activated by inflammatory cytokines, which in turn upregulates hepcidin and interferon regulatory factor 1 (IRF1). IRF1 downregulates SLC7A11, inhibiting GPX4-dependent anti-ferroptotic pathway. (E) The core anti-ferroptotic targets of discussed pathways (cGAS/STING, NF-κB, Nrf2/HO-1 and JAK/STAT) are DHODH, GPX4, and SLC7A11. DHODH reduces CoQ to CoQH2. CoQH2 serves as a potent radical-trapping antioxidant, directly capturing and removing lipid peroxyl radicals on the inner mitochondrial membrane. This pathway runs in parallel with GPX4, offering another line of defense against ferroptosis. The GPX4-dependent anti-ferroptotic pathway and the DHODH-mediated PLOOH clearance pathway in mitochondria are the main characters in this vicious cycle. Created in BioRender. yao, L. (2026) https://BioRender.com/bhfx1mb

4.1 Key molecular pathways related to neuroinflammation-mediated ferroptosis4.1.1 Cyclic GMP-AMP synthase/stimulator of interferon genes

The stimulator of interferon genes (STING) pathway is activated by double-stranded DNA (dsDNA), which activates the expression of interferon (IFN) via interferon regulatory factor 3 (IRF3) (Liu et al., 2015). Beyond its role in promoting inflammation, STING has also been reported to promote ferroptosis in neurons and microglia, mainly via two mechanisms: inhibition of DHODH and suppression of GPX4.

The level of dihydroorotate dehydrogenase (DHODH) was significantly reduced in neurons treated with oxyhemoglobin, a reduction mediated by the CGAS/STING pathway (You et al., 2025). DHODH, a mitochondrial enzyme that suppresses lipid peroxidation in mitochondria and ferroptosis via CoQ (Cao et al., 2025; Mao et al., 2021). The CGAS/STING/DHODH pathway plays a role in not only SAH-induced neuronal ferroptosis but also chronic restraint stress (CRS)-induced hippocampal ferroptosis, resulting in cognitive impairments (You et al., 2025; Zhang J. B. et al., 2025). Inhibition of the CGAS/STING activation and upregulation of DHODH have been shown to reverse depressive-like behaviors, including cognitive and memory defects (You et al., 2025; Zhang J. B. et al., 2025; Tian et al., 2023). Compared with DHODH, the suppression of GPX4 by CGAS/STING has been more well-documented. In a study related to multiple sclerosis (MS), prolonged exposure of neurons to interferon-γ (IFN-γ) induced expression and activation of STING. Rather than canonically activating downstream proinflammatory cytokines, STING promotes autophagic degradation of GPX4, thereby triggering neuronal ferroptosis (Woo et al., 2024). Furthermore, inhibiting the cGAS/STING pathway with RU.521 upregulates the expression of SLC7A11 and GPX4, which mitigates microglial ferroptosis (Sui et al., 2025). Given that both CGAS/STING activation and reduced GPX4 levels have been observed in various neural cells in depression models, their interaction may contribute to ferroptosis and pathological injury in the hippocampus and PFC (Liu et al., 2022; Yuan et al., 2025; Duan et al., 2022; Du et al., 2025).

While the CGAS/STING pathway regulates neuronal and microglial ferroptosis, direct evidence of this connection in depression remains limited. Nevertheless, this regulatory relationship has been demonstrated in various tissues and diseases, suggesting that the mechanisms by which cGAS/STING promotes ferroptosis may be conserved across different cell types and pathological contexts (Gao et al., 2023; Wu et al., 2022; Zhu et al., 2024). Therefore, CGAS/STING may contribute to ferroptosis in depression, given its involvement in multiple CNS disorders, its capacity to induce ferroptosis in various central and peripheral tissues, and importantly, the observation that its inhibition can effectively alleviate depressive-like behaviors. The cGAS/STING pathway thus serves as a critical amplifier of the vicious cycle, linking DNA damage-triggered inflammation to ferroptosis execution.

4.1.2 JAK/STAT

Janus kinase (JAK)/signal transducer and activator of transcription (STAT) is activated by extracellular cytokines such as IL-6 and IFN-γ, and plays a key role in promoting inflammation (Xin et al., 2020). In addition to its inflammatory function, this pathway also promotes ferroptosis, mainly through two downstream effectors: hepcidin and IRF1.

Hepcidin, which can be expressed by CNS cells and transported from the liver, is a downstream target of the JAK–STAT (Zechel et al., 2006; Raha-Chowdhury et al., 2015; Qian et al., 2014). Hepcidin promotes intracellular iron overload via inhibiting iron export through ferroportin blockage and ubiquitination (Kowdley et al., 2021). In the nervous system, the IL-6/JAK/STAT pathway triggers hepcidin expression, as demonstrated in astrocytes after hemoglobin treatment and in neurons via microglial IL-6 release (Qian et al., 2014; Xiong et al., 2016). While some studies have reported anti-inflammatory effects of hepcidin on astrocytes, in other cell types, especially neurons, hepcidin promotes ferroptosis via blocking FPN (Kowdley et al., 2021; Urrutia et al., 2017; Davaanyam et al., 2023), resulting in neuronal loss and contributing to depression (Zhang N. et al., 2025; Farajdokht et al., 2015). The transcription of interferon regulatory factor 1 (IRF1), another downstream factor activated by JAK–STAT, inhibits GPX4 activity by downregulating SLC7A11 (Li et al., 2014; Blaszczyk et al., 2016; Feng et al., 2025; Kong et al., 2021), a component of system XC− that is essential for GSH synthesis (Dixon et al., 2012). Suppressing IRF1 expression with cynaroside, thereby promoting SLC7A11 expression, alleviated ferroptosis in LPS-induced BV-2 cells and reduced hippocampal inflammation and depressive-like behaviors (Zhou Y. et al., 2024). Together, the JAK/STAT/hepcidin and JAK/STAT/IRF1 pathways promote the loss of neurons and microglia, thereby contributing to brain damage and depressive-like behaviors.

4.1.3 NF-κB

The NF-κB pathway can be activated by the binding of Toll-like receptors to ligands such as TNF and IL-1, serving as a master regulator of inflammation (Yu et al., 2020). Although NF-κB is well known for promoting inflammation, it also modulates ferroptosis by suppressing GPX4 and regulating LCN2 expression.

The level of GPX4 is regulated by NF-κB. During ischemic stress, activated NF-κB downregulates GPX4 and system XC− while upregulating TFR1, thereby promoting microglial ferroptosis (Zheng et al., 2024; García-Weber and Arrieumerlou, 2021). Another study related to CIRI also supported the inhibitory effects of NF-κB on GPX4 (Xie et al., 2025). In depression, while direct evidence linking NF-κB and GPX4 remains limited, it has been established that both NF-κB activation and GPX4 suppression contribute to brain injury, including neuronal loss, microglial overactivation, and reduced adult hippocampal neurogenesis (AHN) (Li et al., 2021; Erady et al., 2025).

However, LCN2, a downstream target of NF-κB, exhibits paradoxical effects on ferroptosis—both promoting and inhibiting it in different contexts—adding complexity to the role of NF-κB in this process. LCN2, an acute phase protein secreted by activated astrocytes or neutrophils, has been shown to promote neuronal ferroptosis via downregulation of SLC7A11 in various CNS disease models, including epilepsy and diabetes-associated ischemic stroke (Zhou Z. et al., 2024; Wang et al., 2024). However, NF-κB/LCN2 pathway was reported to inhibit ferroptosis of neurons, which partially protected TBI-induced mice from brain damage (Wang et al., 2025). Although the role of LCN2 in downregulating SLC7A11 has not been extensively studied in depression, LCN2 upregulation and SLC7A11 downregulation were observed in streptozotocin (HFD-STZ)-induced mice, contributing to neuronal loss and cognitive dysfunction (Xie et al., 2023). Furthermore, neuronal ferroptosis and depressive-like behaviors could be elevated by targeting LCN2(Xie et al., 2023; Chen Y. et al., 2023; Yang et al., 2023).

4.2 Key molecular pathways related to ferroptosis-induced neuroinflammation

Ferroptosis not only is a consequence of neuroinflammation but also actively propagates the vicious cycle through the release of damage-associated molecular patterns (DAMPs) and glial activation. DAMPs released from ferroptotic cells—such as lipid oxidation products [hydroxynonenal (4-HNE) and oxidized phospholipids (oxPLs)], HMGB1, and dsDNA—activate the TLR4-MyD88-NF-κB and cGAS-STING pathways, driving the production of proinflammatory cytokines and type I interferons (Wang et al., 2019; Liu et al., 2015; Feng et al., 2024; Imai et al., 2008; Dvorkin et al., 2024). These sequences—from ferroptosis to inflammation mediated by released DAMPs—explain the ferroptosis arm of the vicious cycle. Once initiated, it primes an inflammatory microenvironment for further ferroptotic events. Concurrently, glial cells activated by ferroptosis further amplify inflammation; for instance, iron accumulation in microglia and astrocytes promotes the release of proinflammatory cytokines, exacerbating neuroinflammation and neuronal damage (Ryan et al., 2023; Moro et al., 2025) Through their collective effects, ferroptosis of CNS cells further promotes neuroinflammation.

4.3 Shared molecular pathways

The following contents summarize several key pathways that affect both neuroinflammation and ferroptosis simultaneously.

4.3.1 Nrf2

Nuclear factor erythroid 2 (Nrf2), a protein that regulates genes involved in both ferroptosis and inflammatory pathways, represents a target in depression. Under homeostatic conditions, Kelch-like ECH-associated protein 1 (Keap1) binds to Nrf2 to sequester Nrf2 in the cytoplasm, promoting its degradation via the proteasome (Itoh et al., 1999). Upon stimulation, Nrf2 is released, translocates to the nucleus, and promotes the expression of antioxidant genes, including heme oxygenase 1 (HMOX1), NAD(P)H quinone dehydrogenase 1 (NQO1), superoxide dismutase (SOD), and catalase (