The Evolution of Myelin Oligodendrocyte Glycoprotein Antibody–Associated Disease

Our understanding of central nervous system (CNS) demyelinating diseases has evolved over time, with the recognition of stereotyped clinical phenotypes, improved histopathologic categorization, and more robust assays for identifying CNS-specific autoantibodies. Disorders previously organized by shared clinical phenotype are now better defined by histopathology and clinical biomarkers, including advanced imaging, cerebrospinal fluid (CSF) testing, and serum and CSF autoantibodies.

Myelin oligodendrocyte glycoprotein (MOG) antibody-associated disease (MOGAD) is now identified routinely owing to the development of widely commercially available cell-based assay (CBA) antibody testing. In this review, we examine the history of serologic testing, the natural history of MOGAD and its clinical phenotypes, and our current knowledge of its pathophysiology, and summarize current and potential future therapies.

History of MOG Antibody Serologic Testing

MOGAD is one of the few examples in the history of medicine where the animal model predated our knowledge of the human disease. It took ≈40 years from the initial publication by Lebar et al.¹ in 1976 to the classification of anti-MOG antibodies causing human disease. Their initial article proposed that an immune response to a component of myelin, which they named M2, was responsible for the complement-dependent demyelinating activity they observed in an experimental allergic encephalomyelitis (EAE) guinea pig animal model.¹ Further work demonstrated that M2—what we now refer to as MOG—is found on the surface of myelin and is a potent inducer of immunogenicity in EAE.^1–3 Seven years after the first description of M2, Linnington et al.⁴ identified a mouse monoclonal antibody (8-18C5) that was specific to a glycoprotein found exclusively in the central, but not peripheral, nervous system. Decades later, we now know 8-18C5 as anti-MOG antibody.

MOG is a minor myelin protein, but anti-MOG antibodies play a critical role in animal models in the induction of both an encephalitogenic T-cell response and humoral, antibody-mediated demyelination.^5–7 Early work studying anti-MOG antibody was critical in our foundational understanding of EAE and the pathophysiology of MOGAD, but not until the mid-1990s was the autoimmune response to MOG found to be specific to the immunoglobulin-like N-terminal domain of the protein.⁸ However, conflicting results using linearized or denatured MOG antigen drew into question the role, if any, of anti-MOG antibody pathogenesis.⁹ Subsequent work demonstrated that only anti-MOG antibodies targeting the native conformation of MOG were pathogenic, whereas anti-MOG antibodies targeting linear epitopes of MOG did not cause disease.¹⁰

More than 30 years after the first publication identifying MOG, several groups showed, using CBA testing, that epitope-specific immunoglobulin G anti-MOG antibodies recognizing the native conformational state of MOG were associated with non–multiple sclerosis (MS) demyelinating syndromes.^11,12 These antibodies are not seen in high titers in healthy controls or adults with MS. Low titer antibodies of 1:20 or 1:40, particularly in adults, have a lower positive predictive value for MOGAD than higher titers, and should be interpreted with caution.¹³ Several groups recently published compelling, but early, data showing that a subset of people with MOGAD may have CSF-restricted (ie, seronegative) anti-MOG antibodies.^14,15

Knowledge of Pathophysiology: EAE and Translational Studies

There are limited human histopathologic data on MOGAD. Inflammatory plaques consisting of perivascular and confluent demyelination, composed primarily of a CD4+ T-cell inflammatory response, are most commonly described, as well as intracortical demyelination.¹⁶ The pattern of findings most closely resembles Type II pattern MS plaques. Myeloid cells are seen in both cortical demyelination and in the rim of inflammatory white matter plaques. MOG itself is relatively preserved in samples tested.

The MOG-based EAE models have infiltrates primarily involving the white matter, although both gray matter and retinal disease are seen. The emphasis in MOG-based EAE historically has focused on the role of T cells in disease pathogenesis and propagation.¹⁷ Both CD4+ and CD8+ T-cell responses are appreciated, which may be part of the immunologic bias of the EAE model.¹⁸ T-cell inhibition does not clearly reduce neurodegeneration later in the disease, which in many ways is analogous with the clinical experience in the CNS demyelinating diseases.¹⁹ The role of CD20+ B cells is complex and nuanced. CD20+ B-cell depletion in different MOG-based EAE models ameliorates some and worsens others.²⁰ This is complicated further by CD20+ T cells, which also play an important role in EAE pathogenesis.²¹

Complement pathways play a role in both MOG-based EAE models and MOGAD. Small autopsy and brain biopsy studies suggest a potential role for complement activation, which is supported by elevation in serum complement activation levels in individuals with MOGAD.^16,22 A subset of human anti-MOG antibodies induce complement-dependent pathogenic effects in a murine ex vivo animal model.²³ The role of complement-mediated demyelination is complex, and it is probable that the mechanism is different than what is seen in aquaporin-4 (AQP4)–seropositive neuromyelitis optica spectrum disorder (NMOSD), where complement-mediated demyelination often yields profound disability with poor recovery.^24,25

Whether anti-MOG antibodies are capable of direct independent pathogenicity is unclear. Anti-MOG antibody taken from the serum of people with MOGAD optic neuritis (ON) who had elevated titers appeared to induce demyelination in 2 models of EAE, both with at least some involvement of the complement pathway.²⁶ This finding is not universal in serum obtained from people with MOGAD, so it is possible that these antibodies may be pathogenic through other mechanisms.

Natural History and Clinical Phenotypes

Our understanding of the full spectrum of the clinical phenotype of MOGAD is actively evolving. MOGAD has no known human leukocyte antigen association, and does not have a known predilection for a specific race or ethnicity. MOGAD may have better recovery and faster response to treatment than other acute demyelinating syndromes (ADS).

MOGAD is much more likely to be the cause of an ADS in the pediatric population (upwards of 40% of all ADS cases with ON and acute disseminated encephalomyelitis [ADEM]) compared with the adult population (on the order of 5%).^27,28 As opposed to MS and NMOSD, MOGAD has a female-to-male ratio closer to 1:1. See Table 1 for a comparison of the presentations of MOGAD with those of MS and NMOSD.

High titers of anti-MOG antibodies are not seen in people with MS, and their presence should prompt the clinician to reevaluate the diagnosis, especially if atypical clinical features of MS are present. Because of a lack of formal validated diagnostic criteria to date (the proposed criteria were published in The Lancet on January 24, 2023), it has been up to the clinician to determine the clinical relevance of anti-MOG antibodies, which are routinely sent for many people with ADS.

For organizational purposes, we refer to several typical phenotypes of MOGAD: ADEM, ON, transverse myelitis (TM), and less common manifestations such as cortical encephalitis.

Acute Disseminated Encephalomyelitis

MOGAD commonly presents in children as ADEM. Children presenting with ADEM have a median age of 4 years—a younger age when compared with other phenotypes of MOGAD in children.²⁹ Children younger than 11 years are more likely to present with ADEM with diffuse, bilateral lesions involving the deep gray structures and white matter. Children older than 11 years are more likely to present with no brain lesions with isolated TM or ON, or, when brain lesions are present, the lesions are more likely to have well-demarcated borders.³⁰ Although ≈40% of children with ADEM will have persistent or intermittently positive anti-MOG antibodies over time, the majority of children with persistent antibodies (70%) will still have a monophasic course. In addition, children who become seronegative after initial presentation are less likely to have a relapse, although a relapsing course is not excluded. Because of the high rate of children with MOGAD presenting with a monophasic course of ADEM, most neuroimmunologists do not recommend long-term immunomodulating treatment unless the disease relapses.

Optic Neuritis

In adults and children, ON is the most common presentation of MOGAD. Vision loss is typically severe, but functional recovery is better than that seen in AQP4-seropositive NMOSD. Associated symptoms include periorbital headache, optic disc edema (frequently severe and sometimes associated with hemorrhage), and extensive involvement of the majority of the length of the optic nerve on MRI. Perineural optic nerve sheath enhancement is common.²⁵ ON associated with MOGAD is more likely to involve the anterior optic nerve, whereas AQP4-seropositive NMOSD is more likely to involve the posterior optic nerve. Acute ON associated with MOGAD is more likely to cause substantial retinal nerve fiber layer thickening compared with acute ON associated with MS, which is a potential early biomarker that can be used clinically to differentiate the 2 disorders while waiting for antibody test results.³¹ As opposed to other causes of ON, people with MOGAD are more likely to have mismatch preservation of visual acuity and loss of retinal nerve fiber layer thickness over time on optical coherence tomography.²⁴

Chronic relapsing inflammatory optic neuropathy (CRION) is a potentially distinct demyelinating disease characterized by recurrent unilateral or bilateral ON that is both corticosteroid-responsive and corticosteroid-dependent. Although not all cases are associated with the presence of anti-MOG antibodies, up to 67% of people with CRION test positive for anti-MOG antibodies.³² Although each individual episode of ON tends to respond well to corticosteroids, people with CRION tend to have poor long-term visual outcomes because of the recurrent nature of attacks.

Transverse Myelitis

MOGAD is a common cause of isolated TM, especially in young adults. Although TM in MOGAD frequently presents with longitudinally extensive transverse myelitis similar to that of AQP4-seropositive NMOSD, MOGAD myelitis should be thought of as a distinct disease.³³ Distinguishing among TM associated with MOGAD, MS, or NMOSD is critical to prognostication and treatment. People with MOGAD myelitis are more likely to be age 10 and under, MOGAD myelitis affects male and female individuals equally, and symptoms of myelitis may be preceded by a viral prodrome or vaccination. Ancillary clinical data can help differentiate the disorders. Individuals with MOGAD myelitis are more likely than not to have markedly elevated CSF pleocytosis and to lack oligoclonal bands. Several MRI features may be present, including longitudinally extensive as well as multifocal lesions, involvement of the conus medullaris, and lack of gadolinium enhancement. One feature more specific to MOGAD is isolated involvement of the gray matter of the spinal cord leading to the characteristic H pattern seen on axial imaging of the spinal cord and T2 hyperintense line on sagittal imaging. Because of the frequently seen viral prodrome and gray matter involvement seen on MRI, MOGAD myelitis may be confused for infectious acute flaccid myelitis.

Other CNS Syndromes

In addition to the much more common clinical syndromes mentioned previously, MOGAD also can present with less common clinical phenotypes. In both adults and children, MOGAD can cause a predominantly cortical encephalitis associated with seizures. Anti-MOG antibodies can coexist with other autoantibodies, most commonly anti-NMDA receptor antibodies. When there is coexistence of antibodies, the clinical phenotype tends to resemble the clinical syndrome of the coexistent antibody. Other uncommon syndromes include overlap central and peripheral nervous system inflammation such as radiculitis and other neuritis, which are associated with CNS inflammation (ON, TM, and encephalitis). Although uncommon, MOGAD can cause extensive brainstem and cerebellar lesions.

Prognosis and Natural History of Disease

Prognosis after the incident event and risk of long-term relapse are a function of age. Children 10 and under are more likely to have substantial recovery and lower risk of relapse compared with children age 10 and over and adults (only ~10% of children will have an Expanded Disability Status Scale score ≥3 on follow-up compared with ~25% of adults).³⁴ Many experts recommend a prolonged glucocorticoid taper over several months after acute treatment because of the relatively high risk of relapse upon corticosteroid cessation.³⁵

The risk of relapse in MOGAD is variable, but likely up to 50% of people with MOGAD will not have a relapse. Thus, particularly in children, who have a lower risk of relapse than adults, long-term immunomodulatory therapy often is not started after the incident episode of MOGAD.^34,35 Some clinicians may opt to start long-term therapy in people who have had a particularly severe presentation with substantial disability, but there is no high-quality evidence to guide these decisions. Adults may be at lower risk of relapse if they convert to seronegative antibodies over time, but additional longitudinal data are needed to confirm this. The degree to which antibodies should play a role in making decisions regarding the use of long-term immunomodulatory therapy is unknown. No studies have shown that people with MOGAD are at risk of a primary or secondary progressive MS-like phenotype.

Therapeutics

There are no randomized controlled clinical trials on acute or chronic immunotherapy in MOGAD. Acute therapies include corticosteroids, therapeutic plasma exchange, and intravenous immunoglobulin.³⁶ Whereas the data specific to MOGAD are limited, the suggestion remains that acute therapies are all associated with improved short-term and long-term functional outcomes. There are discrepancies between adult and pediatric populations both in terms of phenotype and utilization of chronic immunotherapy, but prolonged corticosteroid bridges to immunotherapy or prolonged corticosteroid tapers frequently are employed given the risk of early relapse following the initial clinical event.^37,38

Data on chronic immunotherapy to reduce the risk of future relapse are limited to retrospective studies, summarized in Table 2. Long-term immunotherapy largely has been initiated only in people who have already demonstrated a relapsing MOGAD phenotype. Although retrospective studies have numerous limitations in interpretation and translation to clinical practice, there is a plethora of evidence to suggest that immunotherapy reduces the risk of relapse in MOGAD. An early study showed robust response to intravenous immunoglobulin, but subsequent work demonstrated that intravenous immunoglobulin, like other chronic immunotherapies, is likely only partially effective at preventing relapse.^39–41 Navigating the choice of immunotherapy remains a careful discussion between the neurologist and the patient. A variety of therapies targeting both innate and adaptive immunity are underway in late-phase design trials, shown in Table 3.

Summary and Outlook

MOGAD is a relatively newly identified disorder with a wide range of clinical CNS demyelinating phenotypes. International consensus criteria that will emphasize the need for high-quality serologic testing and an expanded range of clinical phenotypes has been long awaited. Monophasic illness is common, but a substantial fraction of people are at risk for relapsing disease. Many individuals with MOGAD may benefit from long-term immunotherapy, and further insight into the precise pathophysiology of MOGAD will be critical in helping identify the most effective treatments to reduce the risk of future neurologic disability. Late-phase clinical trials are underway for therapies with a wide array of mechanisms to reduce the risk of relapse, which in turn will further our understanding of this disorder.