Induction of immune tolerance in NMOSD and MOGAD

Introduction

Neuromyelitis optica spectrum disorder (NMOSD) and myelin oligodendrocyte glycoprotein antibody-associated disease (MOGAD) are autoimmune disorders characterized by inflammatory damage to the central nervous system (CNS).^1,2 While advances in understanding their pathophysiology have illuminated the roles of antibodies such as immunoglobulin G autoantibody targeting aquaporin-4 (AQP4-IgG) and myelin oligodendrocyte glycoprotein (MOG-IgG), therapeutic strategies to induce immune tolerance to these autoantigens remain underexplored. ³ Although MOGAD is classically defined by the presence of pathogenic anti-MOG IgG1 antibodies, a T cell-dependent immunoglobulin subclass, accumulating evidence indicates that MOG-specific CD4⁺ T cells play a central role in disease pathogenesis. ⁴ Experimental autoimmune encephalomyelitis (EAE) models induced by MOG peptide immunization have shown that MOG-reactive CD4⁺ T cells alone are sufficient to drive CNS demyelination, even in the absence of B cells or antibody production. ⁵ In more complex models utilizing full-length human MOG protein, both MOG-specific T and B cells are required; however, B cells primarily serve as antigen-presenting cells (APC) rather than producers of pathogenic antibodies. ⁶ Notably, in transgenic mice expressing both MOG-specific T and B cells but lacking the ability to secrete antibodies, opticospinal inflammation still develops. ⁷ These findings collectively support a model in which T cell-mediated mechanisms can independently initiate and propagate CNS autoimmunity, uncoupled from classical humoral effector functions. Such preclinical observations suggest that a subset of patients currently categorized as double-seronegative NMOSD may, in fact represent a T cell-driven, seronegative form of MOG-directed autoimmunity. ⁴

Immune tolerance is fundamental to preventing autoreactivity, involving central and peripheral tolerance mechanisms to maintain immune equilibrium. ⁸ Dysregulation of these processes underpins the pathogenesis of NMOSD and MOGAD, where autoreactive lymphocytes drive chronic inflammation and tissue injury.^3,8

The immunopathogenesis of NMOSD, primarily mediated by AQP4-IgG, involves astrocyte damage through complement activation and inflammatory cell infiltration. ⁹ In contrast, MOGAD is associated with MOG-IgG targeting oligodendrocytes. ¹⁰ Despite overlapping features, their distinct immune mechanisms necessitate tailored approaches for treatment. ³ For double seronegative (DN) NMOSD cases, where neither AQP4-IgG nor MOG-IgG are detected, a clear autoantigen target further complicates therapeutic strategies, implicating cellular mechanisms such as dysregulated T cells. ¹¹

Current treatments for NMOSD and MOGAD rely heavily on immunosuppression, which could potentially dampen immune responses carrying the risks of infections and malignancies.^3,12 In contrast, antigen-specific tolerance induction represents a promising alternative. Strategies such as T- and B-cell modulation, the use of tolerogenic dendritic cells (DCs), peptide-based vaccines, and nanoparticle (NP)-mediated antigen delivery, aim to restore immune homeostasis while preserving protective immunity. ¹³ Advances in these areas, alongside robust biomarkers for monitoring therapeutic efficacy and disease activity, are critical for optimizing tolerance-based approaches. ⁸

This review explores emerging immunological insights and therapeutic strategies for inducing tolerance in NMOSD and MOGAD, emphasizing the potential of precision medicine to transform management and improve patient outcomes.

Overview of NMOSD and MOGAD

AQP4-IgG-seropositive NMOSD and MOGAD are two recently well-recognized antibody-mediated autoimmune diseases of the CNS.^1,2 For many years, clinical and radiological overlaps between these two disorders led to confusion, often classifying them as variants of multiple sclerosis (MS) due to their shared involvement of the optic nerve and spinal cord. ¹⁴ The discovery of AQP4-IgG and later MOG-IgG clarified this distinction. However, misdiagnosis of these disorders due to misinterpretation and misapplication of diagnostic criteria is still observed in clinical practice.^15,16

A fundamental understanding of antibody testing methods, including the optimal specimen, technique, and potential challenges, is important for diagnosing AQP4-IgG-seropositive NMOSD and MOGAD. The cell-based assay (CBA) in serum is recommended for detecting AQP4-IgG and MOG-IgG, with live CBA presenting certain advantages over the fixed technique.^17–19 In 2007, it was demonstrated that the three-dimensional structure of the MOG protein influences MOG-IgG recognition. Since both glycosylation and the conformation of MOG are critical for accurate recognition, the surface expression of full-length human MOG (typically the α-1 isoform, consisting of 218 amino acids) on HEK293 cells is used to detect MOG-IgG reliably. This method has been instrumental in confirming the presence of MOG-IgG in non-MS.^20,21 While CBA have high specificity, testing a broad and unselected population for MOG-IgG may yield a false-positive rate as high as 28%.^22,23 Moreover, MOG-IgG positivity is rare in adults experiencing their first demyelinating event that resembles MS. ²⁴ Timing of MOG-IgG testing is important,^25,26 as antibody titers fluctuate and may decrease over months from presentation, with some patients subsequently testing negative.^27,28 Recent observational studies suggest that MOG-IgG in cerebrospinal fluid (CSF) may have diagnostic and prognostic utility.^29–33 Close to 40% of patients suspected of having NMOSD who test negative for AQP4-IgG in serum are positive for MOG-IgG using advanced CBA. ³⁴ It became evident that these patients actually have MOGAD rather than NMOSD; thus, enhancing the understanding of various disorders and clarifying the MOGAD phenotype. MOG-IgG was first recognized in children with acute disseminated encephalomyelitis (ADEM), but it is now observed in both adults and children with optic neuritis (ON), transverse myelitis (TM), and brain or brainstem syndromes. ³⁴

MOGAD is primarily associated with demyelination caused by an oligodendrocytopathy, whereas NMOSD involves astrocyte damage due to AQP4-IgG. ³ Cases of concurrent MOG-IgG and AQP4-IgG have been rarely reported, indicating that these antibodies involve distinct immunopathogenic mechanisms.^35–37 A relapsing course is observed in 90% of NMOSD patients and in 43% of MOGAD patients after 8 years. ³⁸ While some patients with MOG-IgG may initially meet NMOSD diagnostic criteria due to overlapping clinical and neuroimaging features, it is now well-recognized that MOGAD is a distinct condition with its own pathophysiology. ² Likewise, it is clear that MOGAD and NMOSD have different risk factors, clinical and radiological findings, and prognostic outcomes, leading to the establishment of separate diagnostic criteria.^1,2

NMOSD typically affects adult women and is classically characterized by severe ON, typically unilateral, with limited visual recovery. However, bilateral involvement may occur, either simultaneously or in a sequential pattern. Optic nerve lesions are usually longitudinally extensive, frequently involving more than 50% of the nerve, and demonstrate a predilection for the posterior segments, including the optic chiasm and optic tract.

Beyond the optic pathway, NMOSD is defined by the presence of longitudinally extensive transverse myelitis, affecting the central gray matter of the spinal cord, and by core clinical syndromes such as area postrema syndrome (manifested by intractable nausea, vomiting, and hiccups), diencephalic syndrome, and symptomatic narcolepsy. ³⁹

In contrast, MOGAD presents in both children and adults, frequently with bilateral or sequential ON, anterior optic nerve involvement with optic disc swelling, and a generally favorable visual prognosis when compared with NMOSD.^1–3,8,9 TM in MOGAD may be longitudinal or short-segment but often shows patchy distribution and preferential involvement of the gray matter, sometimes producing a “H-sign” on imaging. In children, MOGAD is commonly associated with ADEM-like presentations, while in adults, it may manifest with encephalitis, seizures, or brainstem syndromes. While NMOSD is almost invariably relapsing without treatment, MOGAD may follow a monophasic course in a substantial proportion of patients (~50%), although relapses do occur, particularly in those with persistent clinical activity. Recovery from relapses is generally better in MOGAD compared to NMOSD, where attacks are more disabling and cumulative neurological impairment is common.^1,2 The 2023 MOGAD criteria ² outlined three key aspects for diagnosis: (1) core clinical demyelinating events and supporting clinical or magnetic resonance imaging (MRI) features, (2) MOG-IgG and their titers, and (3) exclusion of alternative diagnoses. The panel emphasized that while MOG-IgG serostatus is a critical biomarker, the revised criteria were specifically developed to ensure that the diagnosis of MOGAD is not based solely on antibody status but also includes core clinical and imaging features, thereby minimizing the risk of misclassifying MOGAD patients as seronegative NMOSD or MS. Distinguishing these relapsing neuroinflammatory autoimmune disorders is crucial in clinical practice, as therapeutic strategies differ between these conditions. ³

NMOSD is a spectrum of diseases characterized by recurrent episodes of neuroinflammation and most NMOSD patients test positive for specific AQP4-IgG. ³⁹ NMOSD was previously known as NMO, but the term NMOSD now encompasses the full spectrum of the CNS manifestations, affecting not only the optic nerves and spinal cord but also the brain and brainstem. ¹

The diagnosis of NMOSD was initially based on the presence of AQP4-IgG in serum, with the discovery of these antibodies by Lennon et al. ⁴⁰ in 2004 being pivotal in differentiating NMO from MS. In 2015, NMOSD was defined and stratified according to AQP4-IgG serological status. Additionally, six core clinical characteristics were well-described, and brain and spinal cord magnetic resonance (MRI) findings suggestive or typical of NMOSD were better and well-defined. ¹ However, 20%–30% of patients meeting the 2015 NMOSD diagnostic criteria do not have detectable AQP4-IgG in serum, raising concerns about whether these cases represent the same condition or a distinct entity. ⁴¹

While NMOSD and MOGAD share certain clinical similarities, they are distinct disorders with different immunopathological mechanisms and targets.^1,2 Advances in biomarker identification and diagnostic criteria have enabled the differentiation of these disorders, resulting in more accurate diagnoses and treatment strategies. Ongoing research is important to further understand the pathophysiology of MOGAD and NMOSD, and to enhance therapeutic options for both conditions. Recently, several medications have been developed as molecularly targeted therapies for NMOSD, including complement inhibition, B cell depletion, and interleukin-6 (IL-6) receptor blockade. ¹² These treatments have demonstrated efficacy and safety for NMOSD, and their potential application in MOGAD is also being explored, since there are no approved treatments for this condition.

Underlying immune mechanisms

In NMOSD, IgG1 antibodies targeting AQP4 (a transmembrane protein), called AQP4-IgG, enter from the periphery to the CNS through regions of increased blood–brain barrier (BBB) permeability or injury, or by transcytosis across endothelial cells. ³ AQP4-IgG selectively binds to astrocytic orthogonal arrays of particles (OAPs) of the AQP4 antigen on astrocyte feet, initiating complement-dependent cytotoxicity (CDC) and antibody-dependent cellular cytotoxicity (ADCC), which leads to astrocytic damage, BBB disruption, and subsequent neuroinflammation. ⁴² NMOSD lesions occur where the density of OAPs is highest.^43,44 The first mechanism involves complement cytotoxicity through C5b-9 complex formation (MAC), while the second relies on the sublytic effect of C5b-9 and the dysregulation of astrocyte glutamate transport by C5a-preactivated neutrophils.^45–47 Consequently, inflammatory responses through the recruitment of other innate immune cells such as eosinophils and macrophages, perpetuate the inflammatory process by releasing toxic enzymes and pro-inflammatory cytokines, resulting in oligodendrocyte loss, demyelination, and finally neuron loss. ³ Therefore, AQP4-IgG induces damage based on two major effector functions: First, the binding of AQP4-IgG to OAPs is essential for initiating classical complement pathway activation. This high-avidity interaction is facilitated by the preferential recognition of OAPs over individual AQP4 tetramers. Second, AQP4-IgG engagement induces selective internalization of the M1 isoform of AQP4, a process that disrupts its association with key membrane-bound proteins such as EAAT2, leading to impaired glutamate homeostasis and contributing to excitotoxic damage. ⁴⁸ The involvement of AQP4-specific CD4+ T cells is pivotal for activating B cells and sustaining antibody production, driven in part by HLA-DRB1*03:01-associated immune predisposition. ⁴⁹ Additionally, IL-6, which is a multifunctional cytokine, was found to be elevated in the serum and CSF of NMOSD patients,^50,51 especially during relapses. It contributes to NMOSD pathophysiology by inducing inflammation (supporting plasmablast survival and polarizing CD4+ T cells into Th17 cells), promoting AQP4-IgG production, disrupting BBB integrity, and enhancing proinflammatory T-cell activation.^52,53 Consequently, IL-6 induces astrocyte injury and promotes demyelination. ⁵⁴ It is important to highlight that in NMOSD, the primary target of autoimmune attack is the astrocyte, with demyelination occurring as a consequence of the highly destructive pathogenic process.^55,56

Conversely, MOGAD is characterized by IgG1 autoantibodies targeting the extracellular domain of MOG, which is accessible on the outer surface of oligodendrocytes. ⁵⁷ These antibodies can mediate CDC, ADCC, and antibody-dependent cellular phagocytosis (ADCP). However, current evidence suggests that their pathogenic potential is contingent upon a concomitant pro-inflammatory milieu. In particular, factors such as surface density or conformational presentation that influence antibody binding and immune activation remain incompletely understood. In contrast to NMOSD, complement activation is less prominent in MOGAD lesions, ⁵⁶ and MOG-IgG1 binding appears to occur in a bivalent manner that is suboptimal for C1q recruitment and classical pathway activation,^57,58 though it may enhance Fc receptor-mediated mechanisms such as ADCC and ADCP. ⁵⁸ While complement inhibition is a validated therapeutic strategy in NMOSD,^59,60 its relevance in MOGAD remains uncertain.

Experimental models further support a cooperative mechanism involving both B and T cells. Knock-in mice producing high titers of MOG-specific antibodies fail to develop demyelination in the absence of T cell-mediated inflammation, highlighting the insufficiency of autoantibodies alone to initiate CNS injury.^61,62 This implies that autoreactive B cells may persist in the immune repertoire through clonal ignorance, remaining non-pathogenic under physiological conditions but contributing to disease when inflammatory cues are present. These insights emphasize the need for therapeutic strategies in MOGAD that address both humoral and cellular components of the immune response to effectively mitigate CNS tissue damage. However, MOG-specific T cells are capable of inducing EAE even in the absence of B cells or antibodies.^4,6,63 The heterogeneity of MOGAD, which includes a range of clinical phenotypes like ON, TM, or ADEM, reflects the diverse immunological triggers and mechanisms involved. ³⁴ Current data supporting the pathogenicity of MOG-IgG in vivo is indirect: IgG staining is observed in some MOGAD lesions, and MOG-IgG is found in the CSF of many patients, with titers frequently elevated during attacks. ⁶⁴ Thus, these antibodies may either contribute secondarily to the disease or merely act as a marker of ongoing inflammatory pathology. ⁴

Induction of immune tolerance in NMOSD and MOGAD requires addressing both humoral and cellular immune pathways. In NMOSD, strategies targeting B cells, such as anti-CD20 or anti-CD19 therapies, and complement inhibitors have shown success in reducing relapses by directly mitigating antibody-mediated damage. ¹² However, these therapies do not address the underlying loss of immune tolerance to AQP4, which remains an ongoing challenge. In MOGAD, the focus has shifted toward modulating cellular immunity, particularly targeting MOG-specific T cells that drive inflammation. Tolerance strategies in this context may include antigen-specific approaches, such as tolerogenic DCs or peptide-based immunotherapies, which aim to re-establish immune homeostasis by selectively suppressing pathogenic T and B cell responses.^65,66

Key barriers to achieving immune tolerance include the heterogeneity of immune responses and the lack of robust biomarkers for patients’ stratification.^6,64 In NMOSD, as previously mentioned, the strong association with HLA-DRB1*03:01 suggests that genetic predisposition may play a critical role in tolerance breakdown.^49,67 Identifying linear peptides that comprise an major histocompatibility complex (MHC) class II-restricted AQP4 T-cell epitope sequence linked to a thiol reductase has been shown to generate antigen-specific cytolytic CD4+ T cells capable of selectively eliminating APCs that present this epitope. Furthermore, these cytolytic CD4+ T cells effectively eliminate autoreactive pathogenic T cells targeting the same epitope. ⁶⁸ This approach represents a potential therapeutic strategy for restoring immune tolerance in NMOSD (NCT04629274). In contrast, MOGAD lacks a consistent HLA association, complicating efforts to design targeted tolerance strategies. Moreover, the transient nature of MOG-IgG seropositivity in some patients and the occurrence of relapses despite seroconversion highlight the need for strategies addressing cellular immunity independently of humoral factors. ⁶⁹

Overall, inducing immune tolerance in NMOSD and MOGAD remains an unmet clinical need, requiring innovative approaches that address these disorders’ distinct and overlapping immune mechanisms. Advancing our understanding of the cellular and humoral drivers of autoimmunity will be important for developing targeted therapies that restore immune homeostasis and improve patients’ outcomes.

Current immunotherapeutic landscape in NMOSD and MOGAD

NMOSD and MOGAD are considered heterogeneous disorders, presenting challenges in clinical trial design and efficacy analysis. ⁷⁰ The management goal is to achieve timely diagnosis and early treatment to reduce the frequency and severity of attacks, ¹² as disability is primarily associated with relapses rather than disease progression. ⁷¹

Acute management is crucial in NMOSD and MOGAD, as exacerbations often lead to significant residual disability in NMOSD and a considerable proportion of MOGAD patients may suffer permanent disability because of incomplete recovery from relapses.^34,72

Prompt and intensive treatment is critical in both NMOSD and MOGAD. ⁷³ Although no evidence-based acute management protocols exist, high-dose intravenous methylprednisolone (IVMP, 1 g/day for 3–7 days) is standard. ⁷⁴ Complete recovery after the first attack occurs in 35%–60% of patients treated with IVMP.^74–77 In MOGAD, a tapering course of oral corticosteroids for 3–6 months is recommended to prevent early relapses, with suggested dose reductions to ⩾12.5 ⁷⁸ or ⩾20 mg, ⁷⁹ as reported in Australian and UK cohorts.^27,78,79 While generally well tolerated, the long-term use of corticosteroids remains controversial. ⁸⁰

For NMOSD patients unresponsive to IVMP, plasmapheresis (PLEX), ideally initiated within 5 days, is recommended (5–7 sessions, 1.5 L exchange every other day),^74,81 and similar benefits have been reported in MOGAD. ⁸² PLEX may also be used as a first-line option in severe cases. ⁷⁴ Its therapeutic effect stems from the removal of pathogenic plasma components (immunoglobulins, complement, cytokines) and modulation of both humoral and cellular immune responses.^83–85

Intravenous immunoglobulin (IVIgG) may be considered in NMOSD patients with poor response to IVMP, with retrospective data showing a 50% response rate ⁸⁶ ; some evidence also supports combining IVIgG with IVMP. ⁸⁷ In MOGAD, IVIgG is useful for incomplete recovery or as a long-term option, particularly in children.^81,88

While long-term relapse prevention is standard in NMOSD, the optimal approach in MOGAD remains under debate. Recent consensus guidelines inform NMOSD treatment and switching strategies,^89–91 but similar guidance for MOGAD is still evolving.

Long-term immunosuppressive treatments can be categorized into distinct therapeutic classes:

As of 2025, four therapies, Eculizumab, Ravulizumab, Inebilizumab, and Satralizumab have received regulatory approval for the treatment of AQP4-IgG-seropositive NMOSD. In contrast, no therapies have yet been approved for MOGAD. Nevertheless, several agents (e.g., corticosteroids, MMF, AZA, IVIgG, tocilizumab, rituximab) are employed in clinical practice based on observational data and expert consensus.^{38,80,100–102}

In all clinical trials conducted to date, the primary endpoint has been the efficacy and safety of delaying the time to a new relapse. These therapies have shown significant benefits a preventing subsequent relapses and have demonstrated favorable safety profiles. ¹² However, certain limitations remain, including high treatment costs and increased susceptibility to infections, which may limit their widespread applicability, particularly in resource-limited settings.^12,103 Moreover, these immunosuppressive therapies are not antigen-specific, do not fully suppress underlying inflammatory activity, and fail to eliminate autoreactive immune cell populations, particularly long-lived memory cells. Their efficacy largely relies on sustained immunosuppression, which may compromise protective immune responses against infections and malignancies. ⁸

Consequently, there is growing interest in the development of immune reprogramming strategies aimed at restoring antigen-specific immune tolerance. These emerging approaches seek to promote long-lasting disease remission without impairing the host’s defense mechanisms against pathogens.

Definition and mechanisms of immune tolerance

Immune tolerance is a fundamental physiological process that safeguards the host by preventing inappropriate immune responses against self-antigens, thereby maintaining immune homeostasis and averting autoimmune pathology. ¹⁰⁴ It represents a finely regulated balance between immune activation and inhibition, orchestrated through a multilayered network of cellular and molecular mechanisms. Immune tolerance is broadly classified into central and peripheral tolerance, each operating at distinct stages of immune cell development and activation.

Central tolerance is established during lymphocyte maturation in the primary lymphoid organs, the thymus for T cells and the bone marrow for B cells, where immature lymphocytes expressing high-affinity receptors for self-antigens undergo deletion (clonal deletion) or are rendered non-functional through receptor editing or clonal diversion.^105,106 In the thymus, medullary thymic epithelial cells (mTECs) and DCs present a diverse array of self-antigens, a process governed in part by the autoimmune regulator protein, which facilitates negative selection of autoreactive T cells.^107,108 Similarly, developing B cells in the bone marrow are subjected to receptor editing, deletion, or anergy if they strongly recognize self-antigens.

Despite these stringent mechanisms, central tolerance is not absolute, and autoreactive lymphocytes can escape into the periphery. Peripheral tolerance serves as a critical second checkpoint that regulates mature T and B lymphocytes outside primary lymphoid organs. This is mediated by several non-redundant mechanisms that collectively prevent the activation of autoreactive lymphocytes and maintain immune homeostasis:

In NMOSD, B cells upregulate AQP4 expression after CD40 engagement and present it in an MHC-II context to eliminate AQP4-specific thymocytes. Because AQP4-IgG is an isotype-switched antibody and its production is T cell-dependent, the failure of immune tolerance likely also involves T helper cells. ¹¹⁴ Thymic B cells, rather than mTECs, play a crucial role in tolerating the T cell repertoire against AQP4. These findings support the concept that thymic negative selection of CD40-induced self-antigens in B cells serves as a mechanism to regulate self-destructive T cell–B cell interactions within the systemic immune compartment. Failure of this tolerance mechanism may lead to enhanced T-cell-dependent autoantibody production and the development of overt autoimmune disease. ¹¹⁵

Tregs are present at a reduced frequency in NMOSD patients. ¹¹⁶ A functionally compromised Treg compartment, lacking AQP4-specific T cell clones, may contribute to defective peripheral tolerance, producing pathogenic AQP4-IgG autoantibodies due to insufficient Treg-mediated suppression. ¹¹⁷ Similarly, some studies have reported a reduced proportion of Bregs and decreased IL-10 expression. During acute NMOSD relapses, CD19+CD24hiCD38hi and CD19+CD5+CD38hi Breg cells exhibit significant impairment, accompanied by lower IL-10 levels. ¹¹⁸ The coordinated action of central and peripheral tolerance is essential for maintaining immune homeostasis. Dysregulation of these mechanisms can lead to chronic inflammation and tissue destruction. Understanding these processes is pivotal for developing therapeutic strategies aimed at restoring immune tolerance, which holds significant potential for treating autoimmune and inflammatory disorders while preserving overall immune function. Figure 1 summarizes the main mechanisms involved in the induction of immune tolerance.

OPEN IN VIEWER

Figure 1.

Mechanisms of immunotolerance. Immune tolerance to Self-Ag is primarily established during development, occurring in the thymus for T cells and in the bone marrow for B cells. This process eliminates self-reactive T cells through apoptosis in response to excessive reactivity to self-antigens presented by major histocompatibility complex molecules. Subsequently, positive selection ensures the survival of functionally competent T cells released into the periphery. A similar selection process occurs in the bone marrow to limit B cell autoreactivity. Central tolerance is maintained through receptor editing and apoptosis, significantly reducing the potential for autoimmunity. Receptor editing involves ongoing immunoglobulin light chain gene recombination, leading to secondary rearrangements that modify antigen receptor specificity by replacing one light chain with another. Despite these central tolerance mechanisms, some self-reactive mature T and B cells evade deletion and enter the peripheral lymphoid compartment. Multiple peripheral tolerance mechanisms further regulate immune homeostasis and prevent self-reactive immune cells from mediating tissue damage. These mechanisms include Treg and Breg cells, activation-induced cell death, and the absence of co-stimulatory signals that induce anergy. For clarity, some steps involved in tolerance induction are not depicted in the accompanying schematic.

Importance of immune tolerance in autoimmune diseases

Disruption of either central or peripheral tolerance mechanisms is a hallmark of autoimmune diseases, including MS, systemic lupus erythematosus (SLE), and type 1 diabetes. When immune tolerance fails, autoreactive lymphocytes escape regulatory control, leading to chronic inflammation and progressive damage to specific organs or tissues.^118–120

In disease management, restoring immune tolerance is a cornerstone of addressing the root cause of autoimmune conditions. Unlike traditional immunosuppressive therapies that broadly suppress the immune system, tolerance-inducing approaches aim to target pathogenic immune cells with high specificity while preserving protective immune responses. This precision reduces the risks associated with generalized immunosuppression, such as infections and malignancies, making tolerance-based interventions safer and more sustainable for long-term use. These include the use of tolerogenic DCs, antigen-specific therapies, and biologics that modulate immune checkpoints. ¹⁰⁴ Advances in biotechnology, such as gene editing ¹²¹ and the engineering of regulatory T cells (Treg), ¹²² offer novel tools to precisely control immune responses. Additionally, the development of nanoparticles for targeted antigen delivery ¹²³ and the exploration of mucosal immunotherapy, such as intranasal antigen administration, ¹²⁴ provide innovative methods to induce localized immune tolerance with minimal systemic side effects. To illustrate the clinical translation of these approaches, a phase Ib clinical trial (NCT02283671) evaluating autologous tolerogenic DCs loaded with myelin peptides was conducted in patients with MS, and seropositive and seronegative NMOSD. The trial demonstrated good safety and tolerability, with no serious adverse events reported. Immunological outcomes included significantly increased IL-10 production by peripheral blood mononuclear cells (PBMCs) and trends toward elevated regulatory T cells type 1 (Tr1) cells and reduced T cell proliferation. These findings support the feasibility of antigen-specific tolerance induction in humans, complementing earlier preclinical evidence. By focusing on the root cause of immune dysregulation, these strategies aim to achieve sustained remission and potentially prevent disease activity. Furthermore, these approaches have the potential to transform disease management, offering durable remission, reduced dependence on chronic medications, and improved outcomes and quality of life for patients.

Role of Tregs and Bregs in autoimmune diseases

Tregs and regulatory B cells (Bregs) play critical roles in maintaining immune tolerance and preventing autoimmunity. These cells suppress excessive or aberrant immune responses and promote homeostasis by modulating the activity of effector T cells, B cells, and innate immune cells.^{110–112,125}

Tregs are a subset of CD4+ T cells, distinguished by the expression of the transcription factor FOXP3, which is essential for their development, stability, and suppressive function. Tregs exert their immunosuppressive effects through various mechanisms, including secretion of anti-inflammatory cytokines such as IL-10, TGF-β, and IL-35.^126,127 Additionally, Tregs inhibit immune activation through direct cell-to-cell interactions mediated by molecules like CTLA-4. ¹²⁵

In addition, Tregs can be categorized into different subsets based on their origin and functional properties.^{110–112,128,129}

Bregs suppress inflammatory responses through the secretion of IL-10, IL-35, and TGF-β, as well as through direct interactions with effector cells. ¹³⁰ They modulate T cell activation, suppress DCs maturation, and regulate macrophage polarization toward anti-inflammatory phenotypes. They can be classified into several subsets based on surface markers and cytokine production:

Both Tregs and Bregs represent promising therapeutic targets. Strategies to expand or enhance the functionality of these regulatory cells, such as cell-based therapies or cytokine modulation, are under investigation for their potential to restore immune balance and achieve durable remission in autoimmune diseases.

Limitations in inducing long-term tolerance

Inducing long-term tolerance in autoimmune and inflammatory diseases presents several limitations that require careful consideration. One major challenge is the difficulty in achieving a stable and sustained immune regulatory environment, as immune tolerance induction strategies often face issues of tolerance breakdown over time. This may occur due to the persistence of autoantigens, chronic inflammation, or the activation of compensatory immune responses that counteract the tolerogenic effects. Furthermore, the complexity of the immune system, including the heterogeneous nature of immune responses across individuals, complicates the development of universally effective tolerance-inducing therapies. ¹³⁵

Another limitation is the risk of inducing immune suppression rather than tolerance. ¹²⁰ Tolerance-inducing strategies, particularly those involving immune-modulatory agents such as tolerogenic DCs or biologics, may lead to excessive immune suppression, thereby increasing susceptibility to infections or malignancies. Additionally, the manipulation of immune cells in ways that promote tolerance could inadvertently result in off-target effects or altered immune cell function, potentially worsening disease outcomes or leading to the development of other autoimmune disorders. ¹³⁶

The timing and mode of delivery of tolerance-inducing agents also pose significant challenges. Immune tolerance is often highly dependent on the developmental stage of the immune system, meaning that strategies that are effective at certain stages of immune response may not be applicable later in life or in patients with established disease. ¹³⁷ Moreover, the BBB further complicates treatment by restricting access of therapeutic agents to CNS-resident immune cells and inflamed tissues. Advanced delivery systems, such as nanoparticles or intranasal formulations, are required to effectively transport antigens or tolerogenic agents across the BBB without causing systemic side effects.^138,139

Finally, the heterogeneity of NMOSD and MOGAD, including variations in disease pathogenesis, clinical presentations, and genetic factors, necessitates personalized approaches to tolerance induction. Tailoring therapies to individual patient profiles adds another layer of complexity, requiring a better understanding of the molecular mechanisms underlying disease pathogenesis and immune tolerance. Precision medicine, utilizing biomarkers such as AQP4-IgG, MOG-IgG, or Treg functionality, is essential to tailor treatments and optimize therapeutic outcomes.

Novel approaches targeting B cells, complement pathways, and cytokines

Antigen-specific therapies: AQP4 and MOG antigen-based tolerance induction

The induction of immune tolerance against AQP4 and MOG represents a promising therapeutic strategy to prevent or reverse autoimmune-mediated damage in the CNS. Among the various approaches being investigated, peptide-based therapies aimed at inducing antigen-specific tolerance have garnered significant interest due to their potential to selectively target autoimmune responses while minimizing global immune suppression.

Cell-based therapies

Regulatory T and B cells

Cell-based therapies targeting Tregs and Bregs represent promising strategies for the treatment of autoimmune diseases, including NMOSD and MOGAD. However, the clinical translation of these therapies faces significant challenges, including issues related to cell expansion, safety, and long-term efficacy.

Regulatory T cells

In the context of NMOSD and MOGAD, Tregs have been shown to be dysfunctional or present in insufficient numbers, contributing to the development and progression of the autoimmune response against AQP4 and MOG. Thus, enhancing the number and function of Tregs through adoptive cell therapy has emerged as a potential therapeutic approach for these disorders.^165–167

In animal models of NMOSD and MOGAD, the infusion of ex vivo-expanded Tregs has demonstrated promising results in controlling autoimmune inflammation and preventing CNS demyelination. For instance, in experimental autoimmune NMOSD models, the transfer of AQP4-specific Tregs has been shown to attenuate clinical symptoms, reduce inflammatory infiltrates in the CNS, and prevent the destruction of astrocytes. ¹¹⁷ Similarly, in MOG-induced EAE models, MOG-specific Tregs have been shown to suppress the autoimmune response, reduce the frequency of relapses, and mitigate the severity of demyelination. ¹⁶⁸

The mechanism of action of Treg-based therapies involves the suppression of autoreactive T cells and the promotion of immune tolerance through the production of immunosuppressive cytokines such as IL-10, TGF-β, and IL-35. Tregs can also exert their effects through direct cell-to-cell contact, modulating DCs function and promoting the generation of other regulatory immune cell subsets. ¹⁶⁹

Although preclinical studies have demonstrated the therapeutic potential of Treg-based therapies in autoimmune CNS disorders, several challenges must be overcome for successful clinical translation in NMOSD and MOGAD. One of the primary challenges is the isolation and expansion of Tregs in vitro. While the expansion of Tregs from peripheral blood or tissue samples has been achieved in preclinical studies, the process remains complex and resource intensive. Furthermore, maintaining the stability and suppressive function of Tregs during ex vivo expansion is critical, as expanded Tregs may lose their regulatory properties if cultured under suboptimal conditions or for prolonged periods.^170,171

Another challenge lies in ensuring the specificity of Tregs against AQP4 and MOG without affecting the broader immune system. Strategies such as the use of peptide-loaded DCs or genetic modification to enhance the specificity of Tregs are under investigation but are not yet standardized in clinical practice. ¹⁷²

Moreover, the long-term safety and efficacy of Treg-based therapies remain uncertain. The risk of inducing immune suppression that could lead to opportunistic infections or malignancies must be carefully considered. Additionally, the potential for Tregs to suppress beneficial immune responses, such as anti-tumor immunity, needs to be addressed in clinical trials. ¹⁷¹

Regulatory B cells

In the context of NMOSD and MOGAD, Bregs have been shown to be impaired, contributing to the breakdown of immune tolerance and the exacerbation of the autoimmune response.^173,174 Consequently, Breg-based therapies have been proposed as a means to restore immune balance and reduce autoimmune activity in these diseases.

The mechanism by which Bregs exert their regulatory effects involves the secretion of anti-inflammatory cytokines, which directly inhibit the activation of autoreactive T cells and dampen the pro-inflammatory milieu. Additionally, Bregs can influence the differentiation and function of other immune cells, including DCs and macrophages, to further promote immune tolerance. ¹³³

Despite some promising preclinical data, the clinical application of Breg-based therapies in NMOSD and MOGAD faces several challenges. One of the main issues is the identification and isolation of Bregs from peripheral blood, as these cells represent a rare population and are difficult to distinguish from other B cell subsets. In addition, the expansion of Bregs in vitro is technically challenging, as these cells require specific conditions for optimal growth and function. Furthermore, maintaining the regulatory function of Bregs during ex vivo expansion is a critical consideration, as prolonged culture may alter their phenotype and cytokine production.

A significant challenge in Breg-based therapies is the lack of well-defined markers for identifying functional Bregs. While several markers, including CD19, CD24, and CD38, have been proposed for Breg identification, there is no consensus on the precise phenotype of functional Bregs. This lack of standardization complicates the development of reliable protocols for isolating and expanding Bregs for therapeutic use. ¹¹²

Moreover, the clinical safety and efficacy of Breg-based therapies remain uncertain. The risk of inducing immunosuppressive effects that could result in infections or malignancies is a concern, particularly given the role of B cells in humoral immunity. Additionally, there is the potential for Bregs to inadvertently suppress beneficial immune responses, including anti-infectious or anti-tumor immunity, which must be carefully monitored in clinical trials.

Table 1 includes additional cell-based trials involving mesenchymal stem cells (MSCs) that warrant discussion. In one trial (NCT02249676), the administration of autologous MSCs to refractory NMOSD patients resulted in a significant reduction in disability (EDSS decreased from 4.9 to 4.3, p = 0.02). Similarly, umbilical cord-derived MSCs delivered intravenously and intrathecally (NCT01364246) led to a marked improvement in clinical scores (EDSS from 3.2 to 1.4, p < 0.05). Furthermore, MSCs exhibit paracrine activity through the secretion of immunomodulatory cytokines and trophic factors, and are capable of migrating to sites of injury where they actively participate in tissue repair and regeneration. ¹⁷⁵ These findings suggest that MSCs may exert therapeutic benefits via immunomodulatory, reparative, and neuroprotective mechanisms in both seropositive and seronegative NMOSD cases.

OPEN IN VIEWER

Table 1.

Immune tolerance therapies in clinical development for NMOSD and MOGAD.

Treatment, ClinicalTrials.govIdentifier and clinical phase	Design overview and outcomes	Disease	Intervention and phase of analysis	Main findings and references
Regulatory DCs (NCT02283671) Ib	Open label, Single Group Assignment Primary purpose: Treatment (number of patients with adverse events) Primary outcome: Safety Secondary outcomes: relapses, disability (EDSS), quality of life measures and immunological responses	MS and seropositive or seronegative NMOSD	Ascending dose of tolerogenic DCs loaded with myelin peptides. Current enrollment: 20 (completed)	There were no severe adverse events and was well tolerated. PBMC produced significantly higher levels of IL-10. Trends toward increased Tr1, cells and decreased T cell proliferation in response to AQP4160
Hematopoietic stem Cell transplant (NCT00787722) I/II	Open label, Single Group Assignment Primary purpose: Treatment (survival and PASAT 25-foot walk 9-hole peg test) Primary outcome: Safety Secondary outcomes: attacks, disability (measured by EDSS), QoL scales and immunological responses.	Seropositive NMOSD	Hematopoietic stem cells following pretreatment with IV cyclophosphamide, methylprednisolone, rituximab, and other chemotherapeutic agents. Current enrollment: 13 (completed)	11 out of 13 patients survived more than 5 years post-treatment 175
Autologous mesenchymal stem cells (NCT02249676) II	Open label, Parallel Assignment, placebo-controlled Primary purpose: Treatment (disability) Primary outcome: EDSS change Secondary outcomes: AAR, lesion load and atrophy (MRI), cognition, retinal nerve fiber layer measures and immunological responses.	Refractory seropositive or seronegative NMOSD	Biological: Autologous mesenchymal stem cells Current enrollment: 15 (completed)	EDSS decreased from 4.9 to 4.3 (p = 0.02)
Umbilical cord mesenchymal stem cell therapy (NCT01364246) I/II	Open label, Single Group Assignment, no randomized Primary purpose: Treatment (disability) Primary outcome: EDSS change Secondary outcomes: VEP and lesion load (MRI)	Progressive MS and seropositive or seronegative NMOSD	Biological: combined IV and intrathecal low-dose hUC-MSCs transplantation Current enrollment: 10 (completed)	EDSS decreased from 3.2 to 1.4 (p < 0.05)
CT103A cells I	An ongoing, investigator-initiated, open-label, single-arm Primary outcome: safety and efficacy Secondary outcomes: QoL and disability scales	Refractory seropositive NMOSD	CAR T-cell therapy targeting BCMA Current enrollment: 12 (completed)	All patients had AEs of grade 3 or higher (e.g., hematologic toxic effects) and grade 1 or 2 cytokine release syndrome. 11 patients had no relapse; all patients generally reported improvement in disabilities and QoL outcomes 174
CT103A cells (NCT04561557) I	Open label, Sequential Assignment, no randomized Primary purpose: Treatment (dose-limiting toxicity and incidence and severity of AEs) Primary outcome: Safety Secondary outcomes: pharmacokinetic and Pharmacodynamic, among others.	Refractory seropositive NMOSD and MOGAD	IV infusion of CAR-T cells vs BCMA after receiving lymphodepletion (cyclophosphamide and fludarabine) Current enrollment (estimated): 36 (Active, recruiting) Projected Study Completion: 2027-05-31	N/R
CART cells (NCT05828212) I	Open label, Single Group Assignment Primary purpose: Treatment (dose-limiting toxicity, maximum tolerable dose and incidence and severity of AEs) Primary outcome: Safety Secondary outcomes: AAR, lesion load and Gd+ (MRI), immunological responses, among others	Recurrent or refractory seropositive NMOSD	CD19 CAR-T cells infusion Current enrollment (estimated): 9 (Active, recruiting) Projected Study Completion: 2026-04-30	N/R
RD06-05 CART (NCT06775912) I	Open label, Single Group Assignment Primary purpose: Treatment (Incidence and severity of AEs)	SLE, Systemic Sclerosis, ANCA Associated Vasculitis, Immune-Mediated Myopathies, NMOSD, MS, MG	CAR T-cell therapy administered intravenously after a lymphodepleting therapy regimen consisting of fludarabine and cyclophosphamide Enrollment (estimated): 18 (Active, recruiting) Projected Study Completion: 2027-06-01	N/R
BAFFR CART (NCT06561009) I/II	Open label, Single Group Assignment Primary purpose: Treatment (dose-limiting toxicity and Incidence and severity of AEs) Primary outcome: Safety Secondary outcomes: AAR, lesion load and Gd+ (MRI), immunological responses, among others	Recurrent or Refractory seropositive NMOSD	Anti-BAFFR CART Projected enrollment: 20 (Not yet recruiting) Estimated Study start: 2024-10-01 Projected Study Completion: 2027-10-01	N/R
Universal BCMA-CD19 CART (NCT06633042) I	Open label, Sequential Assignment Primary purpose: Treatment (dose-limiting toxicity and Incidence and severity of AEs) Primary outcome: Safety Secondary outcome: AAR	Refractory seropositive NMOSD	Dose-escalation of universal BCMA-CD19 CART (1.0-4.0 × 106 CAR-T cells/kg) Projected enrollment: 18 (Not yet recruiting) Estimated Study start: 2024-11-25 Projected Study Completion: 2025-12-12	N/R

AAR, annualized relapse rate; AE, adverse effects; ANCA, antineutrophil cytoplasmic antibodies; AQP4, Aquoporin4; BAFFR, B cell activating factor-receptor; BCMA, B cell maturation antigen; CART, chimeric antigen receptor T cells; DC, dendritic cell; EDSS, expanded disability status scale; Gd, gadolinium; hUC-MSCs, human umbilical cord-mesenchymal stem cells; IL-10, interleukin-10; IV, intravenous; MG, myasthenia gravis; MOGAD, myelin oligodendrocyte glycoprotein antibody-associated disease; MRI, magnetic resonance imaging; MS, multiple sclerosis; MSC, mesenchymal stem cells; NMOSD, neuromyelitis optica spectrum disorder; N/R, not reported; PBMC, peripheral blood mononuclear cells; QoL, health-related quality of life; SLE, systemic lupus erythematosus; Tr1, regulatory T cells type 1; VEP, visual evoked potential.

Gene therapy

The application of gene-editing technologies, such as CRISPR/Cas9, has revolutionized the field of immunology and autoimmune disease treatment by offering the potential to directly modify immune cell function and restore immune tolerance. ¹⁸¹

Mechanisms of Gene Editing for Immune Tolerance Gene-editing technologies, including CRISPR/Cas9, TALENs (Transcription Activator-Like Effector Nucleases), and Zinc Finger Nucleases (ZFNs), enable precise modification of the genome by inducing double-strand breaks at specific genomic loci followed by repair processes, either through error-prone non-homologous end joining or homology-directed repair. These technologies can be applied to immune cells to induce tolerance by several mechanisms, including ¹⁸² :

Gene-editing technologies hold promise for restoring immune tolerance, yet several challenges must be addressed before clinical translation. Efficient delivery to immune cells, minimizing off-target effects, and ensuring functional integrity post-editing remain critical hurdles. Ex vivo manipulation and large-scale expansion introduce variability, while immune rejection risks persist even with autologous cells. Long-term efficacy and safety require extensive evaluation, particularly regarding unintended immune dysregulation. Ethical and regulatory frameworks must ensure oversight, especially concerning germline modifications. Advances in precision editing and immune modulation strategies are essential to overcoming these barriers and achieving safe, effective, and durable therapeutic applications.^189,190

Nanoparticles and drug delivery systems

NPs are promising for targeted delivery of antigens and immunomodulatory agents in autoimmune diseases like NMOSD and MOGAD. They overcome challenges of traditional drug delivery, such as poor bioavailability and off-target effects. Engineered from lipids, polymers, or metals, NPs offer precise therapeutic delivery with reduced toxicity, improving treatment outcomes for these conditions. ¹⁹¹ Their applications could include:

Figure 2 summarizes the principal emerging strategies for inducing immune tolerance in NMOSD and MOGAD, while Table 1 provides an overview of the leading immunotolerance therapies currently in clinical development for these disorders.

OPEN IN VIEWER

Figure 2.

Pathophysiological mechanisms and induction of immune tolerance in NMOSD and MOGAD. (1) Peptide-based tolerance induction: Peptide vaccines containing self-antigen epitopes promote immune tolerance by interacting with APCs, thereby modulating T-cell differentiation and regulating the secretion of inflammatory cytokines. (2) Role of tolerogenic DCs: Tolerogenic DCs are crucial for maintaining central and peripheral immune tolerance. They achieve this through mechanisms such as T cell clonal deletion, induction of T cell anergy, and the generation and activation of Treg cells. (3) Breg cell exert immunoregulatory functions through multiple pathways, including the secretion of immunomodulatory cytokines, the release of cytotoxic granzyme B, which induces T cell apoptosis, and the upregulation of FasL, PD-1, and PD-L1. These molecules suppress CD4+ and CD8+ T cell activity and promote the differentiation of Tr1 cells. (4) Targeted elimination of autoreactive B cells: CAR T cells targeting B cell markers such as CD19 facilitate the broad depletion of autoreactive B cells. Alternatively, CAR-modified Treg cells can suppress autoreactive responses without complete B cell depletion. (5) Gene editing technologies, such as CRISPR-Cas9, offer a strategy to restore antigen-specific immune tolerance by engineering DCs into a tolerogenic phenotype, expanding autoantigen-specific Treg and Breg populations and optimizing CAR-T cell function. (6) Synthetic nanoparticles—including polymeric, lipid-based, metallic, and peptide-polymer formulations—can be engineered to co-deliver antigens, antibodies, nucleic acids (DNA/RNA), and immunomodulatory agents in tailored combinations to modulate immune responses.

Conclusion and future perspectives

While substantial progress has been made in understanding the pathophysiology of NMOSD and MOGAD, the mechanisms underlying the induction of tolerance to AQP4 and MOG autoantigens remain unclear. Effective tolerance induction strategies must address autoreactive B cell depletion and the regulation of T cell responses. Regulatory T cells, essential for maintaining immune homeostasis and preventing autoimmunity, are particularly interesting. Antigen-specific Tregs have demonstrated potential in preclinical autoimmune disease models, but selective expansion or activation of Tregs without impairing global immunity remains a challenge.

Emerging strategies, such as peptide-based vaccines or nanoparticle-based antigen delivery systems, aim to induce peripheral tolerance to AQP4 or MOG. These approaches require optimization for efficient antigen presentation, minimal systemic effects, and durable outcomes. Additionally, progress depends on identifying optimal dosing regimens and delivery methods while advancing knowledge of immune pathways regulating tolerance.

A significant hurdle to clinical translation is the absence of reliable biomarkers to monitor treatment efficacy and predict outcomes. Diagnostic markers like AQP4-IgG and MOG-IgG are insufficient for tracking disease activity or therapeutic response. Biomarkers reflecting immune tolerance, such as Treg frequency and function, CNS inflammation resolution, or suppression of autoreactive immune cells, are urgently needed for patient selection, regimen optimization, and trial monitoring.

There are obvious theoretical and practical challenges with these approaches related to their biotechnological feasibility, scarcity of human research data, heterogeneity of NMOSD and MOGAD, prior exposure to immunotherapy, and level of neurological disability.

An important limitation in NMOSD management is the clinical and immunological heterogeneity between seropositive and DN cases. DN-NMOSD patients lack detectable AQP4-IgG or MOG-IgG antibodies, suggesting the involvement of alternative antibodies or non-immune-mediated mechanisms in disease pathogenesis. The absence of defined antigenic targets complicates the development of antigen-specific tolerance strategies. Consequently, treatment remains empirical and phenotype-driven, with acute attacks typically managed as in NMOSD seropositive patients. ¹¹ DN-NMOSD patients may demonstrate lower corticosteroid responsiveness and reduced relapse rates following tapering compared to MOGAD. ²⁰¹ Currently, no relapse-prevention therapies are approved, and conventional immunosuppressants are used off-label. Furthermore, recent clinical trials of eculizumab excluded DN-NMOSD patients, and two randomized controlled trials of satralizumab showed no efficacy in DN-NMOSD individuals, leaving the role of IL-6 and other pathways in this subgroup of patients unresolved.^94,95 However, a critical caveat in interpreting these findings is the limited representation of DN-NMOSD patients within these studies, precluding definitive conclusions regarding the pathogenic relevance of complement and IL-6 pathways in this population.

Developing tolerance-inducing therapies requires robust preclinical and clinical models. While the AQP4-IgG passive transfer model provides insights, it lacks full fidelity to human disease. Advanced models with humanized immune systems or 3D organoids could better simulate CNS pathology, enabling personalized, antigen-specific treatments to restore immune tolerance and reduce CNS damage in NMOSD and MOGAD.