Glial fibrillary acidic protein (GFAP) autoantibody–positive meningoencephalomyelitis is a newly described entity that is typically characterized by a corticosteroid-responsive meningoencephalomyelitis or limited form thereof, accompanied by GFAP-IgG (1,2). Brain MRI often shows perivascular radial enhancement, and imaging of the spine can show longitudinally extensive myelitic lesions (1). Although many of these cases mimic a vasculitic process, cerebral catheter angiography has been normal. Aquaporin-4 and N-methyl-D-aspartate receptor antibodies have been reported to coexist with GFAP-IgG in a third of cases and occasionally, it occurs as a paraneoplastic phenomenon (usually teratoma) (1).
Some patients with GFAP autoantibody–positive meningoencephalomyelitis have been reported to have bilateral optic disc edema (1,2), which is often asymptomatic and was first recognized in this clinical phenotype of encephalomyelitis before discovery of GFAP-IgG (3). The pathogenesis of autoimmune GFAP autoantibody–positive meningoencephalomyelitis and the cause of the optic disc edema are still unknown. We report the ophthalmic details in GFAP-IgG–positive patients seen at the Mayo Clinic to better define the optic disc edema seen in this disease.
This study was approved by the Mayo Clinic Institutional Review Board. Forty GFAP-IgG–seropositive Mayo Clinic patients were identified through our Neuroimmunology Laboratory from January 1, 2000, to December 31, 2016 (1,2). For this study, our inclusion criteria were: A) serum, cerebrospinal fluid (CSF), or both that yielded a characteristic astrocytic pattern of mouse tissue immunostaining with confirmation of IgG reactive with specific GFAPα isoform by cell-based assay (1,2); B) meningoencephalitis or encephalitis; and C) optic disc edema. Fifteen did not have meningoencephalomyelitis or encephalitis and were excluded (14 positive in serum and CSF unavailable and 1 positive in serum but negative in CSF). Eleven patients with GFAP autoantibody–positive meningoencephalitis did not have optic disc edema, although 7 did not have an ophthalmology examination at presentation. We also excluded those with coexisting aquaporin-4-IgG (1 patient) or insufficient clinical information (3 patients).
The medical records of patients were reviewed for the characteristics of optic disc edema, visual acuity, visual fields, optical coherence tomography, fluorescein angiography, MRI findings, lumbar puncture opening pressures, and CSF parameters. Eight of the included patients were reported in previous series on GFAP-IgG (1,2) but did not include detailed information on the ophthalmological findings, which was the focus of this report.
Ten patients were included. GFAP-IgG was detected in CSF, 2 (serum unavailable, 1; serum negative, 1); serum, 3 (CSF unavailable for testing in all 3); or both, 5. The demographics and clinical findings of these patients are shown in Table 1. The median age was 39.5 years (range, 19–74 years) and 60% were men. All patients had meningoencephalitis or encephalitis and 3 also had myelitis. At initial presentation, the optic disc edema was bilateral and symmetric in all patients (Figs. 1A, 2A). Optical coherence tomography showed a thickened retinal nerve fiber layer without any retinal outer layer abnormalities (Fig. 2C). Visual acuity was preserved, and patients were asymptomatic except for 3 patients with transient visual obscurations. On automated perimetry, 2 patients had mild arcuate visual field defects, 1 patient had nonspecific depression, and 1 patient had enlarged blind spots (Figs. 1C, 2E and Table 1).
The opening pressure on lumbar puncture was normal in all patients except for 2 who had mildly elevated pressures of 265 and 298 mm H2O; the median lumbar puncture opening pressure was 144 mm H2O (range, 84–298 mm H2O). All patients had elevated cells and protein in the CSF (Table 1). Fluorescein angiography was available in 1 patient with optic disc edema, which showed leakage from the retinal venules (Fig. 2B). Mild vitritis was noted in this patient and 2 others.
The majority of the patients (90%) had the characteristic radial perivascular enhancement on MRI that can be seen in autoimmune GFAP autoantibody–positive meningoencephalitis (Figs. 1B, 2D and Table 1). The optic disc edema and MRI abnormalities resolved with high-dose intravenous corticosteroid treatment followed by a prolonged oral course. Patient 9 was treated briefly with acetazolamide. Mild optic disc pallor, ganglion cell layer thinning, and persistent arcuate visual field defects were noted in 2 patients after resolution of the disc edema (Fig. 1D, E and Table 1).
GFAP autoantibody–positive meningoencephalitis is a recently described entity characterized by a steroid-responsive meningoencephalomyelitis accompanied by GFAP-IgG (1,2). The clinical presentation of this disorder is quite broad and can range from subacute to chronic encephalitis with or without accompanying meningitis or myelitis. Patients may have characteristic radial perivascular enhancement on MRI, although other enhancement patterns have been encountered (leptomeningeal), and most have an inflammatory CSF. We described 10 patients with bilateral optic disc edema from GFAP autoantibody–positive meningoencephalitis who all had an inflammatory CSF, and 9 of 10 had radial perivascular enhancement on MRI. Most patients were visually asymptomatic, despite the optic disc edema. Two patients had mild optic nerve pallor and arcuate visual field defects after treatment and resolution of the disc edema.
The cause of the optic disc edema remains uncertain. Although it was typically asymptomatic and mimicked papilledema, the opening pressure was normal in the majority of the patients. Only 2 patients in our cohort had a mildly elevated opening pressure, indicating that raised intracranial pressure is unlikely the primary cause of the optic disc edema in the majority of cases. It is possible that elevated CSF protein could have played a role in the elevated intracranial pressure in 2 of our patients. Most patients were evaluated before initiation of therapy, and therefore, steroids did not influence the lumbar puncture opening pressure measurements.
GFAP is found in the retina, especially within the end foot of Müller cells and astrocytes and could be targeted in GFAP autoantibody–positive meningoencephalitis, leading to a breakdown of the retina–blood barrier that contributes to optic disc edema (4). Knockout GFAP experiments in mice have shown locally impaired blood–brain barrier function (5). Fluorescein angiography in 1 patient showed prominent venular leakage. This suggests that the underlying pathogenesis of GFAP autoantibody–positive meningoencephalitis may be a venulitis or at least primarily involves venous inflammation. The characteristic radial perivascular enhancement on MRI also supports a venular process. These findings suggest that the optic disc edema may be a papillitis from an inflammatory vasculopathy as opposed to papilledema from raised intracranial pressure (1). In addition, the fluorescein angiogram had a similar appearance to frosted branch angiitis, which is caused by multiple etiologies including infection, inflammation, and neoplasm, where disc edema is the result of papillitis (6,7).
Hassan et al (8) described a patient with a similar phenotype who presented with bilateral optic disc edema and radial perivascular magnetic resonance enhancement, which was attributed to primary angiitis of the central nervous system because of inflammatory cells infiltrating the walls of small vessels evident on pathology specimens. Similar to our patients, the lumbar puncture showed a normal opening pressure with lymphocytic predominant CSF. In retrospect, we suspect that this may have been a case of GFAP autoantibody–positive meningoencephalitis. Cases with pathology are needed to determine whether GFAP autoantibody–positive meningoencephalitis is primarily an astrocytic disorder or due to venous or arterial inflammation.
Limitations of our study include its retrospective nature and a small sample size, including a very limited number of fluorescein angiograms available to review. Although aquaporin-4-IgG may coexist with GFAP-IgG, we focused on isolated GFAP-IgG cases in this study, as overlapping clinical features would have been difficult to separate. The 3 patients with serum GFAP-IgG seropositivity alone did not have CSF available for testing, but all had the classic clinical phenotype of meningoencephalomyelitis. This is an important finding, as CSF testing of GFAP-IgG has been preferred, given its higher specificity for autoimmune central nervous system disease (1). Not all patients with GFAP autoantibody–positive meningoencephalitis underwent ophthalmic evaluation, and therefore, it is likely that optic disc edema in this condition is more common than reported here.
In summary, the clinical characteristics of GFAP autoantibody–positive meningoencephalitis are still being elucidated. We recommend testing CSF GFAP autoantibodies in patients with unexplained meningoencephalitis, particularly if they have bilateral optic disc edema, accompanying myelitis or radial perivascular enhancement on MRI. The optic disc edema may be due to an inflammatory papillitis affecting the venules as opposed to elevated intracranial pressure. Prospective and pathologic studies will be required to better determine the pathophysiology and frequency of optic disc edema in GFAP autoantibody–positive meningoencephalitis.
STATEMENT OF AUTHORSHIP
Category 1: a. Conception and design: J. J. Chen and E. P. Flanagan; b. Acquisition of data: J. J. Chen and E. P. Flanagan; c. Analysis and interpretation of data: J. J. Chen, A. J. Aksamit, A. McKeon, and E. P. Flanagan. Category 2: a. Drafting the manuscript: J. J. Chen and E. P. Flanagan; b. Revising it for intellectual content: J. J. Chen, A. J. Aksamit, A. McKeon, S. J. Pittock, B. G. Weinshenker, J. A. Leavitt, P. P. Morris, and E. P. Flanagan. Category 3: a. Final approval of the completed manuscript: J. J. Chen, A. J. Aksamit, A. McKeon, S. J. Pittock, B. G. Weinshenker, J. A. Leavitt, P. P. Morris, and E. P. Flanagan.
1. Flanagan EP, Hinson SR, Lennon VA, Fang B, Aksamit AJ, Morris PP, Basal E, Honorat JA, Alfugham NB, Linnoila JJ, Weinshenker BG, Pittock SJ, McKeon A. Glial fibrillary acidic protein immunoglobulin G as biomarker of autoimmune astrocytopathy: analysis of 102 patients. Ann Neurol. 2017;81:298–309.
2. Fang B, McKeon A, Hinson SR, Kryzer TJ, Pittock SJ, Aksamit AJ, Lennon VA. Autoimmune glial fibrillary acidic protein astrocytopathy: a novel meningoencephalomyelitis. JAMA Neurol. 2016;73:1297–1307.
3. Aksamit AJ, Weinshenker BG, Parisi JE. Chronic microglial encephalomyelitis. Poster presentation of the Annual Meeting of the American Neurological Association; October 9, 2012; Boston, MA.
4. Vecino E, Rodriguez FD, Ruzafa N, Pereiro X, Sharma SC. Glia-neuron interactions in the mammalian retina. Prog Retin Eye Res. 2016;51:1–40.
5. Liedtke W, Edelmann W, Bieri PL, Chiu FC, Cowan NJ, Kucherlapati R, Raine CS. GFAP is necessary for the integrity of CNS white matter architecture and long-term maintenance of myelination. Neuron. 1996;17:607–615.
6. Miserocchi E. Frosted Branch Angiitis. Available at: http://www.uveitis.org/docs/dm/frosted_branch_angiitis.pdf
. Accessed August 8, 2017.
7. Walker S, Iguchi A, Jones NP. Frosted branch angiitis: a review. Eye (Lond). 2004;18:527–533.
8. Hassan AS, Trobe JD, McKeever PE, Gebarski SS. Linear magnetic resonance enhancement and optic neuropathy in primary angiitis of the central nervous system. J Neuroophthalmol. 2003;23:127–131.