Immune dysregulation and neuroinflammation in bipolar disorder: Pathophysiological insights and therapeutic perspectives

Abstract

Bipolar disorder (BD) is increasingly associated with immune system dysregulation encompassing both peripheral and central components. Peripheral low-grade inflammation, marked by elevated proinflammatory cytokines and impaired anti-inflammatory responses, contributes to systemic immune imbalance in BD. This peripheral inflammatory state may compromise blood–brain barrier (BBB) integrity, facilitating the entry of peripheral immune mediators into the central nervous system (CNS) and triggering neuroinflammatory cascades. Within the CNS, neuroinflammation is orchestrated primarily by microglial and astrocytic activation, which disrupts neuronal homeostasis and synaptic function. Additionally, oxidative stress acts as a crucial mediator, exacerbating neuronal damage. Based on current findings, this review synthesises evidence linking both central and peripheral inflammation to the pathophysiology of BD, offering a perspective on its underlying biology. A comprehensive understanding of the dynamic interplay between peripheral inflammation and central neuroimmune responses is essential for identifying novel therapeutic targets and developing interventions that effectively address both systemic and CNS components of BD pathophysiology.

Highlights

• •Peripheral inflammation contributes to immune dysregulation in bipolar disorder.
• •Blood–brain barrier dysfunction links systemic and central immune alterations.
• •Microglial and astrocytic activation drive neuroinflammation and neuronal damage.
• •Oxidative stress amplifies inflammatory cascades and impairs neuroplasticity.
• •Targeting immune pathways offer novel therapeutic strategies for bipolar disorder.

Glossary

Bipolar disorder (BD)

A chronic psychiatric condition characterized by alternating episodes of mania/hypomania and depression.

Low-grade inflammation

A persistent, systemic inflammatory state characterized by modestly elevated levels of pro-inflammatory cytokines and acute-phase proteins, often in the absence of overt infection or injury. Unlike acute inflammation, low-grade inflammation is chronic and subtle, but it can influence central nervous system function and modulate the neurotransmission.

Cytokines

Small signalling proteins released by immune cells that regulate inflammation and immunity.

Blood–brain barrier (BBB)

A selective interface formed by endothelial cells and glial structures that controls molecular and cellular traffic between blood and brain.

Microglia

Resident macrophage-like immune cells of the central nervous system that regulate neuronal health and mount inflammatory responses.

Astrocytes

Glial cells that support synapses, regulate neurotransmitter homeostasis, and maintain BBB function.

Oxidative stress

Cellular imbalance between reactive oxygen species production and antioxidant defences, resulting in damage to lipids, proteins, and DNA.

Neuroplasticity

The brain's capacity to reorganize synaptic connections and adapt structurally and functionally to internal and external stimuli.

1. Introduction

Bipolar disorders (BDs) are chronic psychiatric conditions affecting over 1 % of the global population, characterized by recurrent mood episodes, including depressive, manic, and hypomanic periods, often interspersed with phases of euthymia (Grande et al., 2016; Oliva et al., 2025). Moreover, BDs are heterogeneous conditions encompassing distinct clinical subtypes. BD-I is defined by a history of at least one manic episode, namely a period of abnormally elevated, expansive, or irritable mood, accompanied by increased energy, activity, and decreased need for sleep. Manic episodes last at least one week, and cause marked functional impairment, often disrupting occupational and social functioning. On the contrary, BD-II is defined by a history of at least one hypomanic episode and one major depressive episode. Hypomania shares the core features of mania but is less severe, lasts at least four consecutive days, and does not cause significant functional impairment. Major depressive episodes are characterized by persistent low mood, anhedonia, fatigue, sleep and appetite disturbances, and cognitive impairment. Individuals with BD-II often spend more than 80 % of their symptomatic time in a depressive state, which frequently represents the primary driver of functional disability. Across both subtypes, the duration, frequency, and intensity of mood episodes vary widely, contributing to the clinical heterogeneity of BDs and complicating the identification of robust biomarkers and prediction of treatment responses (Berk et al., 2025, Grande et al., 2016, Oliva et al., 2025). Despite advancements in both pharmacological and non-pharmacological treatments, the management of BD remains challenging, as many patients experience inadequate symptom control, adverse side effects, or resistance to available medications (Diaz et al., 2022; Goes, 2023). This therapeutic gap in BD highlights the importance of advancing our understanding of the mechanisms underlying the disorder.

Although several pathophysiological mechanisms have been associated with BD, including neurotransmitter dysregulation, altered neural circuitry, impaired neuroplasticity, metabolic dysfunction, circadian rhythm disruption, and alterations in gut microbiota (Dalkner et al., 2021; Ortega et al., 2023; Yan et al., 2023), inflammation has gained prominence as a central mediator that may shape these dysregulations (Pereira et al., 2021). At the molecular level, inflammation is primarily mediated by cytokines and chemokines, small signalling proteins secreted by immune cells such as macrophages, T lymphocytes and endothelial cells. Key cytokines include interleukin-6 (IL-6), interleukin-1 beta (IL-1β) and tumour necrosis factor-alpha (TNF-α), which coordinate a cascade of immune responses, ranging from neutrophil activation and B cell proliferation to acute-phase protein synthesis and increased vascular permeability (Grebenciucova and VanHaerents, 2023).

Importantly, these inflammatory processes interact with multiple other pathways implicated in BD. Cytokine-mediated inflammation can modulate neurotransmitter systems, altering serotonin, dopamine, and glutamate signalling, and can affect neural circuit function, particularly within prefrontal-limbic networks involved in mood regulation (Miller et al., 2013; Saggu et al., 2025). Chronic inflammation may also impair neuroplasticity, reducing neurogenesis and dendritic remodelling, and contribute to metabolic dysregulation, promoting insulin resistance, obesity, and dyslipidemia, which are frequently observed in patients (Calkin, 2019; Pinzi et al., 2025). Together, these findings underscore the multifaceted impact of immune activation on both brain and systemic physiology, highlighting the need for an integrated perspective.

Here, we propose an integrative model that moves beyond domain-specific summaries to elucidate how peripheral and central immune alterations converge in the pathophysiology of BD. Specifically, this review seeks to (i) summarize current evidence on peripheral immune alterations, including cytokine imbalance and oxidative stress; (ii) elucidate the mechanisms by which these peripheral processes affect central nervous system function through BBB disruption, endothelial dysfunction, and immune-to-brain signaling; and (iii) examine the downstream neuroinflammatory responses involving microglial and astroglial activation, redox imbalance, and impaired neurotrophic support. By integrating these levels within a unified systems immunology framework, the review aims to clarify how peripheral–central immune interactions shape neuroprogression and clinical heterogeneity in BD, and to identify potential targets for personalized and mechanism-based therapeutic strategies.

Despite clear interconnections across immune and neurobiological domains, most systematic and narrative reviews have addressed inflammatory dysregulation in BD by considering peripheral biomarkers, blood–brain barrier (BBB) alterations, glial activation, and oxidative stress in isolation. To date, no work has articulated a cohesive mechanistic framework integrating these processes within a systems immunology perspective.

In this sense, the review advances the field by explicitly integrating peripheral and central immune mechanisms within a unified systems immunology framework that connects molecular, cellular, and clinical levels, offering mechanistic hypotheses relevant to biological stratification and the development of targeted therapeutic strategies. Accordingly, here, we also discuss the relevance of these mechanisms as potential targets for adjuvant therapeutic strategies. A better delineation of these pathways, despite the heterogeneity among studies, may ultimately guide the development of novel interventions modulating both peripheral and central immune responses.

2. Peripheral inflammation in bipolar disorder

Peripheral inflammation has emerged as a key component of the pathophysiology of BD, offering novel insights into disease mechanisms. This is supported by findings showing that elevated levels of circulating cytokines, including IL-6, TNF-α, and IL-1β, have been consistently reported during acute manic and depressive episodes, and are thought to influence neuroplasticity, synaptic pruning, and neurotransmitter regulation, thereby impairing cognitive performance (van den Ameele et al., 2017; Misiak et al., 2018). Neuroimaging studies further show that higher peripheral cytokine levels are associated with structural brain alterations, including reduced cortical thickness and compromised white matter integrity (Benedetti et al., 2016). While these cytokine elevations often normalise during euthymia, mild immune alterations may persist, reflecting a more chronic, trait-like dysregulation (Sayana et al., 2017). Importantly, rather than representing isolated molecular changes, these findings suggest altered coordination within immune signaling networks, consistent with a systems immunology perspective in which cytokines act as interdependent nodes within dynamic regulatory circuits.

Among the cytokines that have been studied in BD, TNF-α emerged as the main central player. This pleiotropic cytokine, primarily secreted by monocyte-derived cells such as macrophages, plays a key regulatory role in inflammatory networks and is involved in essential biological processes including cell signalling, differentiation and apoptosis (Bradley, 2008; Parameswaran and Patial, 2010).

Increased levels of TNF-α have been consistently reported in individuals with BD, particularly during acute mood episodes, reflecting a state of peripheral immune activation. Notably, these elevations are detectable even in the early stages of illness, including first episodes, where they coincide with the onset of cognitive and functional impairment (Chakrabarty et al., 2019). This suggests that inflammatory disturbances may be intrinsic to the disease process, rather than merely a consequence of illness chronicity. However, this view is tempered by the inconsistency of findings across studies, as others failed to observe significant cytokine alterations in the early stage of disease (Solmi et al., 2021).

Importantly, in BD-I patients, these elevations persist into the later stages of illness, indicating ongoing activation of the TNF signalling pathway (Kauer-Sant’Anna et al., 2009). Such chronic upregulation may reflect a cumulative inflammatory burden, an impaired resolution of immune responses, or a progressive sensitisation of inflammatory circuits following recurrent mood episodes. Corroborating this hypothesis, the levels of TNF-α soluble receptor sTNFR1 have been shown to correlate positively with both illness duration and the number of depressive episodes, suggesting a potential role in tracking disease progression (Doganavsargil Baysal et al., 2019).

Together, these findings support the evidence that TNF-α and its soluble receptors are not only elevated across mood states and illness stages, but may also reflect underlying biological processes driving the course of BD.

In line with findings on TNF-α, elevated levels of IL-6 and its soluble receptor sIL-6R have also been consistently reported in individuals with BD, particularly during acute mood episodes. IL-6 is a pleiotropic cytokine predominantly produced by macrophages, with a key role in initiating and modulating immune responses (Grebenciucova and VanHaerents, 2023). Multiple studies showed increased IL-6 concentrations during both manic and depressive phases, which tend to return to baseline during remission (Lu et al., 2019; Luo et al., 2016). Beyond mood symptomatology, elevated IL-6 has been associated with clinical severity, including higher scores on the Young Mania Rating Scale (YMRS) (Luo et al., 2016), and with neurobiological alterations such as cognitive impairment (Barbosa et al., 2018). In a cross-sectional study of euthymic adults with BD-I, serum IL-6 levels negatively correlated with global cognitive performance and specific domains, including language, visuospatial abilities, and orientation/attention, while positively correlating with the number of hospitalizations (Ríos et al., 2025). Importantly, IL-6 significantly predicted cognitive outcomes across multiple domains and hospitalization frequency. Moreover, meta-analytic and cross-sectional evidence further shows that whereas CRP and TNF-α elevations are largely restricted to acute episodes, IL-6 remains elevated even in euthymia, suggesting it may function as a trait marker (Solmi et al., 2021). Collectively, these findings suggest that IL-6 may not only act as a state peripheral marker of inflammation during mood episodes but also contribute to the broader pathophysiological processes underlying the disorder. Accordingly, these data fit a network-based model in which IL-6 contributes to maintaining system tone rather than simply indicating transient activation.

Interferon-gamma (IFN-γ), a key pro-inflammatory cytokine involved in Th1-mediated immune responses (Alspach et al., 2019), has been implicated in the immunopathophysiology of BD. Primarily produced by natural killer (NK) cells and T lymphocytes, IFN-γ plays a pivotal role in immune regulation, antiviral defence, and macrophage activation (Alspach et al., 2019). Some studies have reported elevated IFN-γ levels in individuals with BD, suggesting a role in sustaining peripheral immune activation (Argue et al., 2025; Wu and Zhou, 2024). However, findings remain inconsistent as other studies, including meta-analyses, failed to detect significant differences or have even reported lower IFN-γ levels in BD compared to controls (Modabbernia et al., 2013; Su et al., 2002). Nevertheless, this variability may reflect state-dependent oscillations within immune network dynamics rather than contradictory evidence.

Additional evidence for the role of IFN-γ signalling in affective disorders arises from clinical conditions where interferons, such as interferon-alpha (IFN-α) for chronic hepatitis C or interferon-beta (IFN-β) for multiple sclerosis, are used therapeutically, and their use has been associated with the onset or worsening of depressive and manic symptoms (Capuron and Miller, 2011; Lesh et al., 2018). Therefore, these psychiatric side effects provide indirect support for the hypothesis that heightened interferon activity contributes to mood dysregulation.

Importantly, the impact of these markers may extend beyond immune activation to encompass neuroendocrine modulation. Indeed, similar to TNF-α and IL-6, IFN-γ has been shown to activate the hypothalamic–pituitary–adrenal (HPA) axis, resulting in increased cortisol secretion (Pariante and Lightman, 2008). While this response represents an adaptive reaction to acute stress, chronic elevation of inflammatory cytokines may drive persistent hypercortisolemia, glucocorticoid receptor desensitization, and impaired negative feedback regulation. Such alterations have been consistently linked to core features of BD, including emotional dysregulation, cognitive impairment, and increased vulnerability to mood episodes (Belvederi Murri et al., 2016). Thus, these cytokines may contribute to BD not only by sustaining peripheral immune dysregulation but also by altering several neuroendocrine pathways, which are crucial in shaping mood and cognitive functioning. In systems terms, these cross-system feedbacks exemplify how immune and neuroendocrine pathways are coupled components of a shared regulatory architecture, where persistent perturbation in one domain (e.g., immune signaling) can propagate through others (e.g., stress response and neuroplasticity).

CRP, a sensitive peripheral marker of low-grade inflammation, has also been implicated in the pathophysiology of BD. CRP is mainly produced by hepatocytes but can also be synthesised by macrophages and endothelial cells. It contributes to inflammation by activating the complement cascade, promoting phagocytosis and apoptosis, and inducing nitric oxide production. Notably, CRP production is downstream of pro-inflammatory cytokines such as IL-6 (Sproston and Ashworth, 2018), upregulating systemic inflammatory responses. Like data from cytokines modulation in BD, also blood levels of CRP have been found significantly elevated during manic episodes (Cunha et al., 2008) and, to a lesser extent, also during depressive phases in BD-I patients (De Berardis et al., 2008). Euthymic patients generally exhibit reduced or normalised CRP levels compared to mood episodes, although not always reaching levels observed in healthy controls (Fernandes et al., 2016; McNamara and Lotrich, 2012). Importantly, CRP levels have been found reduced by pharmacological treatment and associated with symptom severity as measured by YMRS scores in BD-I patients (Uyanik et al., 2015). Comparable findings have been consistently reported in major depressive disorder (MDD), where meta-analyses show elevated CRP levels, particularly in drug-naïve and first-episode patients, positively correlating with both central inflammatory markers and depressive symptom severity. Interestingly, while both disorders share this peripheral low-grade inflammatory signature, comparative studies suggest that CRP elevations are more pronounced during acute bipolar episodes than in unipolar depression, converging to similar levels during remission. This evidence supports the view that inflammation represents a transdiagnostic mechanism across mood disorders, with disorder-specific trajectories and magnitudes (Poletti et al., 2024). IL-4 is one of the most investigated anti-inflammatory cytokines in BD. It is produced primarily by Th2 cells and mast cells and promotes the differentiation of naïve T cells into Th2-type cells while inhibiting Th1-driven pro-inflammatory responses (Luzina et al., 2012). In BD-I, elevated IL-4 levels have been reported during manic episodes compared to both depressive phases and healthy controls, suggesting a phase-specific compensatory response aimed at counteracting heightened systemic inflammation (Fiedorowicz et al., 2015; Ortiz-Domínguez et al., 2007). Notably, IL-4 concentrations were also higher in BD-related mania than in major depressive disorder (MDD, highlighting its potential utility in differentiating depressive episodes across diagnostic categories (Lu et al., 2023). However, findings have not been entirely consistent: some studies report reduced IL-4 levels in BD, which negatively correlate with mood severity scores, indicating that insufficient IL-4 may impair the resolution of inflammation and contribute to mood instability (Wu and Zhou, 2024). These divergent results may reflect a dynamic and context-dependent modulation of IL-4 based on mood phase, illness stage, and individual variability in immune reactivity. This duality underscores the complexity of IL-4 signalling in mood disorders, where compensatory responses may themselves become dysregulated and eventually not be beneficial as they could exacerbate immune dysregulation itself.

IL-10, one of the most potent anti-inflammatory cytokines, is primarily produced by regulatory T cells (Tregs), monocytes, macrophages, and dendritic cells (Iyer and Cheng, 2012). It acts by suppressing the expression of pro-inflammatory cytokines such as TNF-α, IL-6, and IL-1β, thereby attenuating systemic immune activation (Burmeister and Marriott, 2018). Elevated IL-10 levels have been observed from the earliest stages of BD, including during the first manic episode and subsequent remission, suggesting a role in the resolution and recovery from inflammatory activation (Remlinger-Molenda et al., 2012). In contrast, more advanced stages of the disorder are characterised by a progressive decline in IL-10 levels (Kauer-Sant’Anna et al., 2009), evaluated in BD-I, suggesting a loss of anti-inflammatory regulation and greater vulnerability to the neurotoxic effects of chronic inflammation. Notably, IL-10 levels tend to normalise with chronic lithium treatment, possibly contributing to its long-term immunomodulatory effects by suppressing pro-inflammatory cytokines, protecting neuronal integrity, and restoring immune balance (Damri and Agam, 2024). This observation underscores the broader relevance of anti-inflammatory cytokines in both the pathophysiology and clinical monitoring of mood disorders. In this context, IL-4 and IL-10 have been investigated as potential biomarkers to distinguish BD from MDD, with multivariate feature selection and machine learning approaches demonstrating moderate discrimination between the two disorders (sensitivity 62 %, specificity 66 %) (Poletti et al., 2024). Therefore, systematic monitoring of IL-10, and potentially IL-4, could not only inform treatment response but also provide prognostic insight and aid in differentiating bipolar from unipolar depressive episodes. Recent evidence further suggests that the link between inflammation and BD may be partly mediated by metabolic dysregulation. In particular, obesity and related metabolic abnormalities can influence inflammatory tone and, in turn, disease expression. In a large cohort, elevated cytokine concentrations in BD lost statistical significance after adjusting for body mass index, underscoring a complex interplay between inflammatory pathways, metabolic status, and clinical phenotype (Knight et al., 2023).

Overall, this immune imbalance in BD supports the presence of a chronic low-grade inflammatory state, potentially contributing to both symptom expression and illness progression. Their persistent increase during acute episodes and, in some cases, during euthymia, supports a role as both state- and trait-related biomarkers, with implications for diagnosis, prognosis, and disease monitoring. Notably, some inconsistencies across studies likely reflect differences in mood state, illness duration, medication status, comorbid conditions, sample size, and measurement methodologies, rather than a true absence of inflammatory alterations. Indeed, these differences can result in phasic variability for specific cytokines. TNF-α, sTNFR1, sIL-2R, IL-6, and IL-1RA show state-dependent changes, whereas IL-4 and IL-10 do not (Modabbernia et al., 2013), highlighting how specific inflammatory markers may be differentially influenced by clinical state. Additional sources of heterogeneity include variations in sampling conditions (fasting status, time of day), sporadic control for confounders such as BMI, smoking, physical activity, blood pressure, and alcohol consumption, as well as differences in medication regimens and assay types. Moreover, sensitivity and subgroup analyses indicate that some of these factors, such as assay method or plasma versus serum choice, can influence specific cytokine outcomes underscoring the importance of cautious interpretation (Munkholm et al., 2013). These contradictions suggest that inflammatory and immune alterations in BD are unlikely to follow a uniform trajectory. Rather, distinct immune phenotypes appear to exist. Furthermore, their interplay with metabolic networks and neuroendocrine systems, particularly HPA axis dysregulation, indicates that immune dysregulation in BD arises not from a single pathway but from multi-level interactions spanning immune, metabolic, and neuroendocrine systems, eventually influencing mood regulation and cognitive functioning, offering potential targets for personalized therapeutic strategies. This supports the view that peripheral immune alterations are not merely epiphenomena, thereby linking systemic inflammation to core clinical manifestations such as affective instability and cognitive impairment. Moreover, these cytokine shifts should not be interpreted in isolation but as part of a broader immune network imbalance that may propagate through vascular and glial pathways.

2.1. Peripheral oxidative stress

An additional contributor within the inflammatory cascade of BD is peripheral oxidative stress (Serban et al., 2025). Accumulating findings indicated that patients with BD display systemic redox imbalance, reflected by impaired antioxidant defences and increased oxidative damage. Specifically, reduced activity of key antioxidants, including superoxide dismutase (SOD), catalase, glutathione peroxidase (GPx) and glutathione (GSH) has been consistently reported in plasma (Andreazza et al., 2008; Jiménez-Fernández et al., 2021). In parallel, elevated levels of malondialdehyde (MDA), 4-hydroxynonenal (4-HNE) as well as 8-hydroxy-2′-deoxyguanosine (8-OHdG) have been observed in peripheral blood samples from individuals with BD (Jiménez-Fernández et al., 2021). Notably, oxidative imbalance in peripheral blood has been linked to greater illness severity, cognitive impairment, and reduced treatment response (Jiménez-Fernández et al., 2021). Pro-oxidant markers tend to rise during mania (Jiménez-Fernández et al., 2021; Lima et al., 2022; de Sousa et al., 2014). Lithium appeared to exert beneficial effects on peripheral redox balance, as shown by reductions in lipid peroxidation along with antioxidant enzyme activity following treatment in BD-I and BD-II (de Sousa et al., 2014).

Importantly, peripheral oxidative stress may reinforce systemic inflammation. Excessive production of reactive oxygen species (ROS) promotes the activation of redox-sensitive transcription factors, including NF-κB and AP-1, upregulating the expression of pro-inflammatory cytokines such as IL-1β, IL-6, and TNF-α (Maes et al., 2011). Similarly, sustained inflammation enhances ROS production through the induction of NADPH oxidase (NOX) and inducible nitric oxide synthase (iNOS), thus causing a self-amplifying loop between inflammation and oxidative stress.

Taking this further, loss of immune regulation may act as a key upstream driver of central neuroinflammation, contributing to neuronal dysfunction, impaired neuroplasticity, and cumulative disease burden. Understanding how peripheral immune alterations interface with CNS processes is, therefore, essential for developing interventions that target both clinical symptoms and the mechanisms underlying illness progression.

3. Central neuroinflammatory mechanisms in bipolar disorder

Recent insights have drawn increasing attention to neuroinflammation as a potential mediator of immune dysregulation effects on brain function in BD.

Numerous studies have indeed identified elevated levels of central proinflammatory cytokines in individuals with BD, particularly during acute mood episodes (Chaves-Filho et al., 2024). Söderlund et al. (2011) observed increased IL-1β and decreased IL-6 levels in the cerebrospinal fluid (CSF) of BD-I and BD-II patients compared to controls, suggesting dysregulated central immune signalling (Söderlund et al., 2011). Isgren et al. (2017) further reported that even euthymic BD-I and BD-II patients exhibit elevated CSF concentrations of IL-8, MCP-1 (CCL2), YKL-40 (CHI3L1), and neurofilament light chain (NF-L), all markers of neuroinflammation and neuronal injury (Isgren et al., 2017). Meta-analyses have confirmed these alterations. Niu et al. (2019) reported significantly elevated levels of inflammatory markers in the CSF of individuals with BD (Niu et al., 2019). While IL-8 showed a trend toward increased concentrations, findings for IL-6 remained inconsistent, likely due to methodological and sample heterogeneity across studies.

These inflammatory mediators can exert detrimental effects on neurogenesis and neural plasticity in key brain regions such as the prefrontal cortex and hippocampus (Han and Ham, 2021). Among them, IL-1β has been shown to play a central role in suppressing adult neurogenesis. Koo and Duman (2008) demonstrated that hippocampal progenitor cells in the subgranular zone express IL-1β receptors, and their activation leads to reduced cell proliferation (Koo and Duman, 2008). Interestingly, this effect was reversed when IL-1β signalling was blocked. In rodent models, systemic administration of lipopolysaccharide significantly promotes regional neurodegeneration through elevated IL-1β and TNF-α levels (Pienaar et al., 2025). Similarly, Keohane et al. (2010) found that exposing differentiating hippocampal neural progenitor cells to TNF-α resulted in impaired neurogenesis (Keohane et al., 2010).

In addition, central immune abnormalities have been linked to greater illness severity, as reflected by a higher number of mood episodes, longer disease duration, and increased medical comorbidities in patients with BD (Muneer, 2016).

Taken together, these findings underscore the pivotal role of neuroinflammation in the pathophysiology of BD, that contribute to neurodegeneration, impaired neurogenesis, and ultimately disease progression.

3.1. Blood-brain barrier dysfunction

A key mechanism linking peripheral immune activation to central neuroinflammation may involve BBB dysfunction, which has emerged as a critical feature in the pathophysiology of BD (Wakonigg Alonso et al., 2024; Zhao et al., 2022). In line, from a systems immunology perspective, the BBB may act as a key interface mediating the crosstalk between peripheral and central immune compartments, translating systemic inflammatory states into local neural responses. The BBB is a highly selective, semipermeable interface formed by endothelial cells, tight junction proteins, pericytes, and astrocytic end-feet, maintaining CNS homeostasis by regulating the passage of molecules and immune cells between the periphery and the brain (Kadry et al., 2020). In BD, compromised BBB integrity may allow peripheral immune cells and inflammatory mediators to enter the CNS, thereby amplifying neuroinflammation, driving neuronal dysfunction, and contributing to disease progression and cognitive decline (Che et al., 2024; Millett Caitlin et al., 2025, Millett et al., 2025). Emerging biomarker studies support the presence of BBB abnormalities in BD. Elevated levels of matrix metalloproteinase-9 (MMP-9), intercellular adhesion molecule-1 (ICAM-1), claudin-5, and the astrocyte-derived protein S100B have all been linked to impaired BBB function and disease progression (Wakonigg Alonso et al., 2024). Among these, MMP-9 is consistently increased in BD, where it promotes extracellular matrix degradation and tight junction disruption (Reininghaus et al., 2016, Zhao et al., 2022. Similarly, ICAM-1, a marker of endothelial activation, is elevated in BD-I and BD-II and may facilitate leukocyte adhesion and transmigration across the endothelium (Schaefer et al., 2016). Altered claudin-5 expression has been associated with increased BBB permeability and heightened psychiatric vulnerability in BD models (Greene et al., 2020). Collectively, these findings highlight BBB dysfunction as a pivotal pathway through which peripheral immune activation can exacerbate central pathology in BD.

Notably, increased BBB permeability has been associated with more severe cognitive symptoms in BD patients, particularly with deficits related to attention, executive functioning, and memory (Depp et al., 2012, Knorr et al., 2024, Millett et al., 2025). These cognitive symptoms may result from the increased entry of peripheral immune mediators and neurotoxic substances into the CNS, which trigger sustained neuroinflammatory processes and synaptic alterations. To confirm that, a study by Zhu et al. (2025) demonstrated that individuals with BD and elevated inflammatory profiles exhibit significant alterations in angiogenesis-related signalling pathways, which may contribute to BBB dysfunction. Specifically, this subgroup exhibits increased mRNA expression of angiopoietin-2 and SERPINE1, alongside reduced protein levels of angiopoietin-1, angiopoietin-2, and the TEK receptor (Zhu et al., 2025). Angiopoietin-1 normally functions to stabilise the vascular endothelium and prevent permeability, whereas angiopoietin-2 is typically upregulated during inflammation and acts in opposition, promoting vascular destabilisation. The imbalance observed in these BD patients suggests a suppression of angiopoietin signalling in the midbrain, potentially impairing vascular stability and integrity. Such molecular alterations support the hypothesis that dysregulated angiogenesis, particularly under inflammatory conditions, plays a role in compromising the BBB in BD.

Although findings remain partly inconsistent, growing evidence indicates that vascular and endothelial dysfunction are present in BD not only during acute episodes but also during remission, supporting the notion that BBB abnormalities may represent core features of the disorder (Zhao et al., 2022). Importantly, BD-I and BD-II patients with higher levels of BBB leakage were found to exhibit a more chronic course of illness, suggesting that BBB disruption contributes to disease progression by sustaining neuroinflammation and cumulative neuronal injury (Kamintsky et al., 2020). Moreover, improvement in BBB disruption with lithium has been reported in both in vitro studies and animal models of ischemic-stroke and stress-related mood disorders (Ji et al., 2021; Taler et al., 2021). Although these are not direct models of BD, they remain informative, as they capture convergent mechanisms, such as neuroinflammation and BBB dysfunction, that are also implicated in BD pathophysiology, thereby providing valuable insights into lithium potential protective effects on BBB integrity. Moreover, recent evidence suggests that BBB may itself impact lithium response; since over half of BD patients fail to respond to lithium, variability in BBB function could partly explain treatment resistance (Luo et al., 2021). Investigating lithium effects on BBB may therefore open new chances for personalized treatment strategies.

Taken together, current evidence indicates that BBB dysfunction in BD may represent a central mechanism linking peripheral inflammation to long-term illness progression, cognitive decline and treatment response. Accordingly, BBB could be considered as a potential therapeutic target in BD. Thus, future research should assess how BBB alterations shape illness trajectory and treatment responses, opening the possibility that BBB-targeted strategies may complement current therapeutic interventions.

3.2. Glial dysfunction

Building on evidence of BBB disruption in BD, the observed neurovascular alterations may further exacerbate neuroinflammatory processes, particularly through mechanisms involving microglial and astrocytic activation in response to BBB disruption.

Following a damaging insult, the CNS mounts a neuroinflammatory response largely mediated by the activation of microglia, the brain's resident macrophages (Gao et al., 2023). Once activated, microglia can adopt distinct functional phenotypes. The M1 phenotype is characterised by the release of pro-inflammatory cytokines and reactive oxygen species, contributing to tissue injury and neurotoxicity. In contrast, the M2 phenotype promotes anti-inflammatory signalling, tissue repair, and neuroprotection. This dynamic polarisation highlights the dual role of microglia as both drivers of pathology and facilitators of recovery, depending on the balance between M1 and M2 states (Cherry et al., 2014). While transient M2 activation supports tissue repair and homeostasis restoration, sustained M1 activation results in the release of ROS and inflammatory cytokines and chemokines, such as TNF-α, IL-6, IL-1β, IL-12, and CC chemokine ligand (CCL) 2 (Palasz et al., 2023). Aberrant microglia–neuron interactions have been increasingly implicated in the pathophysiology of BD, with converging evidence from post-mortem studies supporting a role for microglial dysregulation (Pinto et al., 2018a). Specifically, increased expression of inflammatory and glial activation markers, including IL-1β, IL-1R, MyD88, NF-κB, GFAP, iNOS, c-Fos, and CD11b, has been reported in the frontal cortex of individuals with BD, pointing to a sustained neuroinflammatory state (Rao et al., 2010). Moreover, an imbalance between M1 and M2 polarisation of microglia seems to contribute to the progression of BD (Pinto et al., 2018b). In vivo PET imaging studies further corroborate these findings, revealing focal microglial activation in the hippocampus of euthymic BD-I patients (Haarman et al., 2014). Additionally, CSF from euthymic BD-I and BD-II patients showed elevated levels of microglial activation markers such as MCP-1 and YKL-40 (Jakobsson et al., 2015).

Astrocytes, the most abundant glial cells in the central nervous system, are essential regulators of neuronal and network function. They contribute to synaptic support and plasticity, maintain glutamate uptake and recycling to prevent excitotoxicity, provide key metabolic substrates to neurons, and play a pivotal role in sustaining BBB integrity and cerebral homeostasis. Emerging evidence implicates astrocytic dysfunction in the pathophysiology of BD. In contrast to their neuroprotective A2 counterparts, A1 astrocytes have neurotoxic properties, as they promote neuronal and oligodendrocyte death (Liddelow et al., 2017). Indeed, a compromised BBB integrity exposes astrocytes to peripheral inflammatory mediators, which trigger reactive and potentially maladaptive glial responses. Inflammatory A1 astrocytes may in turn compromise BBB function by releasing matrix metalloproteinase-8 (MMP-8) and pro-inflammatory cytokines, thereby increasing vascular permeability and contributing to disease progression (de Rus Jacquet et al., 2023). Eventually, astrocytic dysfunction and BBB disruption appear to be mutually reinforcing. Supporting this, Toker et al. (2018) reported increased expression of astrocyte-associated genes in BD patients (Toker et al., 2018). In this context, glial gene expression analyses in the dorsolateral prefrontal cortex and anterior cingulate cortex identified pronounced sex-related differences in BD, with males exhibiting significantly higher expression of glia-associated genes, including astrocytic markers (Zhang et al., 2020). These transcriptional alterations may reflect sex-specific astrocytic responses to inflammatory stimuli, potentially contributing to the differential clinical presentation and course of BD across sexes.

Collectively, these findings underscore the intricate interplay between microglia and astrocytes in BD, positioning glial activation as a central hub within the immune–brain network. Through mechanisms involving BBB disruption, persistent inflammatory signaling, and maladaptive glial responses, this network amplifies and perpetuates peripheral immune signals within the CNS, ultimately sustaining neuroinflammatory processes. Elucidating the temporal sequence and causal pathways of these alterations through longitudinal and mechanistic studies will be critical to clarify their role in illness progression. Future longitudinal and mechanistic studies may therefore inform the development of targeted interventions aimed at attenuating neuroinflammation, ultimately modifying disease trajectory.

3.3. Oxidative stress

As seen in the periphery, oxidative stress plays a key role also in the CNS (Șerban et al., 2025), linking immune dysregulation to neuronal damage and reinforcing the interplay between systemic and central inflammatory mechanisms. In the CNS, a sustained activation of microglia and astrocytes drives an excessive release of ROS and RNS, including nitric oxide (NO), which can overwhelm endogenous antioxidant defences such as glutathione, superoxide dismutase (SOD), and catalase (Chen et al., 2020; Ishihara and Itoh, 2023). High levels of oxidative damage markers, such as 4-hydroxynonenal (4-HNE), a byproduct of lipid peroxidation, and 8-hydroxy-2′-deoxyguanosine (8-OHdG), an indicator of oxidative DNA damage, have been reported in the anterior cingulate cortex of individuals with BD (Wang et al., 2009). Importantly, these alterations are accompanied by elevated expression of neuroinflammatory markers. Although the literature remains limited, studies by Andreazza et al. (2013) have reported both oxidative and inflammatory alterations in the prefrontal cortex, supporting the hypothesis of intertwined neuroinflammatory and oxidative pathways contributing to the pathophysiology of BD (Andreazza et al., 2013). Complementary evidence from neuroimaging supports the presence of central oxidative dysregulation in BD. Magnetic resonance spectroscopy studies have consistently reported reduced levels of glutathione, a key antioxidant, in the anterior cingulate cortex of BD-I patients, particularly during mood episodes. This suggests that oxidative imbalance may function as a state-dependent marker of disease activity (Versace et al., 2014). These findings underscore the relevance of redox imbalance in brain circuits involved in affective regulation. From therapeutic perspectives, these insights have fuelled growing interest in antioxidant-based interventions. For instance, N-acetylcysteine (NAC), a glutathione precursor, have shown promise in clinical trials, with evidence of improvements in depressive symptoms and functioning in BD-I and BD-II patients (Berk et al., 2008; Samuni et al., 2013). Although results remain heterogeneous, oxidative stress further reinforces this interconnected system, serving both as a downstream consequence of inflammation and as a driver of ongoing immune dysregulation, eventually highlighting the potential of targeting redox imbalance as an adjunctive treatment strategy.

4. Integrative framework: A systems immunology model of bipolar disorder

Across the studies reviewed, immune dysregulation in BD emerges not as a series of independent findings but as a distributed network disturbance involving peripheral and central components of the immune system. Rather than isolated inflammatory signals, these data collectively suggest a multi-level regulatory failure spanning cytokine production, BBB integrity, glial activation, and oxidative homeostasis.

In this systems immunology framework, BD pathophysiology can be conceptualized as a dynamic interaction between the peripheral immune system, the BBB, and central glial networks. Peripheral low-grade inflammation initiates a cascade of molecular events comprising altered cytokine signaling, increased oxidative stress, and disruption of tight-junction proteins, that compromises BBB function. This barrier dysfunction facilitates the entry of immune mediators into the CNS, thereby amplifying local neuroinflammation. Activated microglia and reactive astrocytes, in turn, release additional pro-inflammatory and oxidative molecules, perpetuating peripheral immune activation.

This loop provides a mechanistic explanation for the oscillating, yet persistent inflammatory profile observed in BD, as well as its association with neuroprogressive changes and treatment resistance. Such a network-based interpretation integrates findings from peripheral biomarkers, neuroimaging, and cellular studies into a unified conceptual model that views BD as a disorder of immune–neural communication and systemic homeostatic imbalance.

From this perspective, the diverse biological alterations reported across studies can be understood as interconnected parts of a common multi-systemic process, rather than as independent pathophysiological domains. Recognizing BD through this systems-level lens may facilitate the identification of immunological subtypes and guide the development of targeted interventions aimed at restoring immune and neural equilibrium.

5. Targeted immune-based interventions in bipolar disorder

In this context, converging lines of evidence raise the question of whether the dysregulated inflammatory pathways observed in BD are already being targeted therapeutically. Conventional mood-stabilisers, such as lithium and valproate, not only regulate mood symptoms but also exert immunomodulatory effects: lithium, for example, has been shown to reduce pro-inflammatory cytokines (including IL-2 and IL-6) from peripheral blood leukocytes and to modulate glial activation in both in vitro and clinical studies (Damri and Agam, 2024). Moreover, as shown by Wu et al., lithium treatment in patients with BD modulates immune cell populations and transcriptomic profiles, leading to down-regulation of inflammation-related pathways and highlighting a mechanistic link between mood stabilization and immune modulation (Wu et al., 2019).

Beyond conventional agents, both preclinical and clinical studies have explored adjunctive anti-inflammatory compounds, in particular Non-Steroidal Anti-Inflammatory Drugs (NSAIDs), pioglitazone, omega-3 fatty acids, and biologics targeting TNF-α, IL-6R or IL-1β, which might offer precision immunotherapy to BD patient sub-groups with elevated baseline inflammatory burden. Building on these findings, several clinical studies have evaluated adjunctive anti-inflammatory strategies in BD. N-acetylcysteine (NAC) has consistently shown benefits over placebo, improving symptom severity, response, and remission rates, while exhibiting a favourable safety profile. NSAIDs and coenzyme Q10 (CoQ10) have also demonstrated efficacy in some trials, though the evidence is less robust and further high-quality studies are needed. By contrast, biologics targeting specific cytokines, such as TNF-α inhibitors, have yielded mixed results, underscoring the potential need to stratify patients according to their inflammatory profile to optimise therapeutic outcomes (Xu et al., 2023).

Overall, these studies suggest that stratifying patients using peripheral blood biomarkers, neuroimaging correlates of neuroinflammation and clinical phenotypes may allow tailored interventions, enhancing efficacy while minimising unnecessary exposure.

6. Conclusions

This review integrates evidence across molecular, cellular, and systemic levels, highlighting the intricate interplay between peripheral immune dysregulation and central neuroinflammation in the pathophysiology of BD. Rather than a purely neurocentric condition, BD increasingly appears as a multi-system disorder, in which peripheral and neural networks are tightly interconnected. Cumulative evidence suggests that low-grade peripheral inflammation may not merely reflect an epiphenomenon of mood episodes but acts as a peripheral trigger and actively contributes to CNS vulnerability. These peripheral disturbances interact with the integrity of the BBB, allowing peripheral mediators and immune cells to enter the CNS, amplifying neuroinflammatory responses. Within the brain, these processes trigger microglial activation and astrocytic reactivity, shifting glial populations toward pro-inflammatory, neurotoxic phenotypes and further promoting oxidative stress, ultimately forming a self-reinforcing loop that disrupts neuronal plasticity, neurogenesis, and synaptic homeostasis, thereby sustaining affective instability and cognitive decline. Moreover, this contribution cannot be fully understood through a simplistic pro-versus anti-inflammatory framework. Cytokine signaling in BD is highly dynamic and context-dependent, reflecting feedback interactions among immune cells, glia, endocrine systems, and metabolic pathways, with functional outcomes shaped by illness phase, cellular milieu, and feedback regulation. Adopting a systems immunology perspective therefore provides a more integrated view of how peripheral immune alterations can drive central neurobiological changes. Once central, these inflammatory signals appear to trigger maladaptive glial responses and oxidative stress, collectively impairing neurogenesis, synaptic plasticity, and neuronal homeostasis (Fig. 1). The integration of these dysregulations within a systems framework supports the concept of a neuro-immune axis, where maladaptive feedback between peripheral and central immune networks leads to instability across mood-regulating circuits. Importantly, acute episodes correspond to transient pro-inflammatory surges, while residual dysregulation persists during euthymia as a trait-like vulnerability. By explicitly conceptualizing these processes as components of a multi-level immune system, this framework provides a unifying explanation for the coexistence of peripheral inflammation, BBB dysfunction, glial activation, and oxidative stress.

Fig. 1.

Interplay between peripheral and central immune dysregulation in bipolar disorder

Peripheral immune cells such as macrophages and T lymphocytes release pro-inflammatory cytokines (TNF-α, IL-6, IL-1β) and acute-phase proteins like CRP. These changes are accompanied by oxidative stress, reflected in increased levels of reactive oxygen species (ROS).

The blood–brain barrier represents the critical interface between the periphery and the brain. Dysfunction of the blood–brain barrier, characterized by altered expression of MMP-9, ICAM-1, Claudin-5, and S100B, allows peripheral cytokines and immune cells to penetrate the central nervous system.

Within the brain, microglia shift toward a pro-inflammatory M1 phenotype and astrocytes become reactive, releasing cytokines and ROS. These processes further compromise blood–brain barrier integrity and amplify neuroinflammation. Neurons exposed to these insults show oxidative damage, leading to reduced synaptic plasticity and progressive neurodegeneration.

This vicious cycle ultimately contributes to the clinical manifestations of bipolar disorder, including mood instability, cognitive impairment, illness progression, and treatment resistance.

Despite these insights, a major limitation is the heterogeneity across studies. Variability in patient characteristics, mood states, illness stages, medication status, and methodological approaches complicates the interpretation of findings and the identification of robust biomarkers. This heterogeneity may mask consistent patterns, but it also highlights the potential existence of clinically relevant subgroups, suggesting that personalized approaches may be critical for future therapeutic strategies. Accordingly, these findings carry potential therapeutic implications, and interventions targeting immune dysregulation may exert benefits by modulating underlying neurobiological processes.

In addition, emerging data suggest that specific inflammatory profiles may differentiate BD from other psychiatric conditions such as MDD (Poletti et al., 2024). For instance, studies have found that IL-10 levels are higher in BD compared to MDD, whereas IL-1β tends to be more elevated in MDD, suggesting that these conditions may exhibit partially distinct inflammatory profiles. (Goldsmith et al., 2016). A recent comparative review reported that MDD is more frequently associated with anti-inflammatory cytokines IL-4 and IL-10, whereas BD showed stronger links with CRP (Poletti et al., 2024). These differential patterns further underscore the relevance of inflammation not only as a shared mechanism in affective disorders but also as a potential marker for diagnostic stratification and targeted interventions.

In summary, from a systems immunology standpoint, BD emerges as a disorder of multi-layer immune failure in which peripheral and central immune dysregulation, BBB alterations, glial dysfunction, and oxidative stress are interdependent and mutually reinforcing, contributing to illness progression. Future longitudinal and mechanistic studies are essential to delineate causality, clarify the interactions among peripheral and central processes, and inform precision medicine approaches that address both systemic and CNS components of the disorder.

CRediT authorship contribution statement

Floriana De Cillis: Writing – original draft, Investigation, Conceptualization. Veronica Begni: Writing – original draft, Visualization, Investigation. Ilari D'Aprile: Writing – original draft, Investigation, Conceptualization. Giulia Petrillo: Writing – original draft, Investigation, Conceptualization. Marco Andrea Riva: Writing – review & editing. Annamaria Cattaneo: Writing – review & editing, Supervision.

Data availability statement

No new data were created or analysed in this study. Data sharing is not applicable to this article.

Funding

This work was supported by the EarlyCause Project (grant no.848158) to A.C., and by the Italian Ministry of Health to A.C. (Ricerca Corrente).

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

No data was used for the research described in the article.