Clinical features of three cases of neurosarcoidosis with a positive aquaporin-4 antibody requiring differentiation from neuromyelitis optica spectrum disorder

Abstract

We report three cases of neurosarcoidosis presenting with optic neuritis or myelitis that required differentiation from neuromyelitis optica spectrum disorder (NMOSD) due to positive aquaporin-4 antibody (AQP4-Ab) results. Positive AQP4-Ab findings were identified using a cell-based assay (CBA) in one case and enzyme-linked immunosorbent assay (ELISA) in two cases. The positive results obtained by CBA were considered to represent either latent positivity or false positives, whereas all positive results obtained by ELISA were regarded as false positives. All cases showed elevated soluble interleukin-2 receptor levels in both the serum and cerebrospinal fluid, as well as enlarged mediastinal and hilar lymph nodes. One patient responded poorly to biological monotherapies including ravulizumab. Positive AQP4-Ab results caused by neurosarcoidosis should be considered in cases with features that are atypical for NMOSD.

Translated Abstract

Introduction

Sarcoidosis is an autoinflammatory disease of unknown etiology. Characterized by granuloma formation and subsequent fibrosis, the condition primarily affect the lungs, skin, liver, and joints¹⁾. Neurosarcoidosis occurs in approximately 5% of systemic sarcoidosis cases²⁾ and exhibits heterogeneous clinical manifes­tations. Granulomatous inflammation may involve the meninges, cranial nerves, brain, spinal cord, and peripheral nerves, leading to diverse clinical presentations of the disease³⁾. Diagnosing neurosarcoidosis can be difficult since approximately 10–19% of patients exhibit no symptoms of systemic extraneural inflam­mation⁴⁾. Neuromyelitis optica spectrum disorder (NMOSD) is an autoimmune condition affecting the central nervous system and characterized by the presence of pathogenic autoantibodies, specifically aquaporin-4 antibodies (AQP4-Ab). Lesions typically present as optic neuritis and longitudinally extensive myelitis involving three or more vertebral segments⁵⁾.

Neurosarcoidosis can also present with optic neuritis or myelitis, conditions often requiring differentiation from NMOSD⁶⁾⁷⁾. In this report, we describe three patients with positive AQP4-Ab results who were ultimately diagnosed with neurosarcoidosis instead of NMOSD.

Case Report

Case 1 (supplemental figure 1)

The patient was a 60s man with a history of uveitis. Three weeks prior, he developed gradually progressing bilateral vision loss. He was referred to our department because of the absence of worsening uveitis. Neurological examination revealed loss of bilateral light reflexes and visual loss in the right index and left manual values. Blood tests showed normal soluble interleukin-2 receptor (sIL-2R) levels at 550 ‍U/ml (normal range: 121–613 ‍U/ml) and negative results for antinuclear antibody, Sjögren syndrome-A/B (SS-A/B) antibodies, and myeloperoxidase/proteinase 3-antineutrophil cytoplasmic antibodies (MPO/PR3-ANCA). The enzyme-linked immunosorbent assay (ELISA) for AQP4-Ab, conducted by an external laboratory (SRL, Inc., Tokyo, Japan), and the cell-based assay (CBA) for myelin oligodendrocyte glycoprotein antibodies (MOG-Ab) both yielded negative results. Cerebrospinal fluid (CSF) analysis revealed a markedly elevated cell count of 57/‍μl, comprising predominantly mononuclear cells (56/‍μl) and a single polymorphonuclear cell (1/‍μl). The CSF protein level was significantly elevated at 160.0 ‍mg/dl, and the sIL-2R level was also elevated at 153 ‍U/ml (reference range: <50 ‍U/ml). In contrast, the immunoglobulin G (IgG) index remained within normal limits at 0.69. Whole-body computed tomography (CT) did not reveal any abnormalities. Similarly, MRI of the brain and spinal cord showed no pathological findings. However, orbital MRI demonstrated bilaterally enlarged peripapillary lumens and areas of high signal intensity, consistent with a diagnosis of bilateral optic perineuritis (Fig. 1A, B). Based on these findings, the patient was diagnosed with idiopathic optic perineuritis and initiated on intravenous methylprednisolone (IVMP) at a dose of 500 ‍mg/day for three consecutive days, followed by oral prednisolone (PSL) at 30 ‍mg/day. The treatment improved visual acuity to 0.5 in both eyes. Subsequently, the diagnosis was revised to NMOSD after a positive AQP4-Ab result on CBA, which was determined by an outsourced laboratory (Cosmic Corporation Co., Ltd., Tokyo, Japan). Treatment with tacrolimus 2 ‍mg/day was added, leading to further improvement in visual acuity to 0.8 in the right eye and 1.0 in the left eye. Subsequently, PSL was tapered to 5 ‍mg/day for maintenance. However, 6 months later, the patient’s visual acuity decreased to 0.08 in the right eye and 0.05 in the left eye. Orbital MRI confirmed bilateral optic perineuritis, indicating the recurrence of NMOSD. IVMP 1,000 ‍mg/day was administered for 3 days, followed by ravulizumab 3,300 ‍mg every 8 weeks, which decreased the CH50 from 60 ‍U/ml (normal range: 25–48 ‍U/ml) to 15 ‍U/ml. Consequently, the patient’s visual acuity improved, and PSL and ‍tacrolimus were tapered off. Nevertheless, 10 months later, the patient’s visual acuity decreased again to 0.07 in the right eye ‍and 0.06 in the left eye. Plasma exchange and IVMP 1,000 ‍mg/day for 3 days were administered over two courses, resulting in improved visual acuity. As the patient experienced a relapse during treatment with ravulizumab, the possibility of therapeutic failure due to a genetic polymorphism in C5 was considered. In view of the high disease activity, it was deemed necessary to switch to another biologic agent. The patient was transitioned from ravulizumab to inebilizumab for recurrence prevention. However, bilateral visual deterioration recurred four months later. Subsequent laboratory investigations revealed markedly elevated levels of sIL-2R at 1,699 ‍U/ml and angiotensin-converting enzyme (ACE) at 24.0 ‍U/l (reference range: 8.3–21.4 ‍U/l). CT imaging demonstrated enlargement of the mediastinal and hilar lymph nodes (Fig. 2A, B). Histopathological analysis of mediastinal lymph node tissue obtained via transbronchial biopsy revealed non-caseating epithelioid granulomas (Fig. 3A), leading to a revised diagnosis of probable neurosarcoidosis. Repeat testing for AQP4 antibodies using a CBA remained negative. Infliximab was administered alongside PSL at 15 ‍mg/day. The patient did not experience recurrence at the last follow-up.

Fig. 1  Orbital and spinal cord magnetic resonance imaging.

Short tau inversion recovery images in Case 1 show bilaterally enlarged peripapillary lumens and high signal intensity (A, B). In Case 2, T₂-weighted imaging revealed a longitudinally extensive spinal cord lesion extending from C2 to C8 (C). Gadolinium-enhanced T₁-weighted imaging shows contrast enhancement from C5 to C6 (D). Axial T₂-weighted images revealing a high signal intensity involving the entire spinal cord at the C6 level (E), with corresponding gadolinium enhancement on T₁-weighted images (F). Axial T₂-weighted images showing a high-signal area extending from the central spinal cord to the dorsal region at the Th6 level in Case 3 (G).

Fig. 2  Chest computed tomography scan.

Enlarged mediastinal and hilar lymph nodes observed in Case 1 (A, B), Case 2 (C, D), and Case 3 (E, F).

Fig. 3  Pathological findings.

Biopsy of a hilar lymph node reveals non-necrotizing epithelioid granulomas on hematoxylin–eosin (H&E) staining in Case 1 (A). Spinal cord biopsy showing non-caseating epithelioid granulomas on H&E staining in Case 2 (B). H&E: hematoxylin-eosin

Case 2 (supplemental figure 2)

The patient was a 60s man with a history of hypertension and glaucoma in the right eye. He experienced abnormal sensations in his fingers and both lower limbs for 2 years and gradually developed lower limb muscle weakness. Neurological exami­nation revealed weakness of the distal muscles of the left upper extremity (manual muscle testing [MMT] grade 3), weakness of the proximal muscles of the left lower extremity (MMT grade 4), generalized hyperreflexia of limb tendons, abnormal sensation and hyperalgesia in both palms, and hyperalgesia below the Th8 level. Blood tests were negative for antinuclear antibodies, anti-SS-A antibodies, and MPO/PR-3 ANCA but showed an elevated sIL-2R level of 1,176 ‍U/ml and a positive AQP4-Ab result of 5.8 ‍U/ml on ELISA. The ACE level was normal at 17.4 ‍U/l. CSF analysis revealed an elevated cell count of 6/‍μl (5/‍μl mononuclear cells and 1/‍μl polymorphonuclear cells), protein concentration of 180 ‍mg/dl, and sIL-2R level of 121 ‍U/ml. However, the IgG index was normal at 0.67. Spinal cord MRI demonstrated a large, high-signal intramedullary lesion extending from C2 to C8, with contrast enhancement from the left to the ventral sides at C5 and C6 (Fig. 1C–F). Whole-body CT revealed mild enlargement of the bilateral hilar and mediastinal lymph nodes (Fig. 2C, D). A spinal cord biopsy was performed, and histopathological examination revealed a non-desmoplastic granuloma (Fig. 3B), leading to a diagnosis of spinal neurosarcoidosis. Furthermore, AQP4-Ab retesting by CBA was negative, which indicated that the ELISA result was a false positive. IVMP was administered at 1,000 ‍mg/day for 3 days, followed by PSL at 60 ‍mg/day and methotrexate (MTX) at 7.5 ‍mg/week. While the abnormal sensations improved, muscle weakness persisted.

Case 3 (supplementary figure 3)

The patient was a 70s man who underwent pacemaker implantation for complete atrioventricular block 6 months before. Two weeks before admission, he experienced muscle weakness and abnormal sensations in his left leg, which later spread to both hands. As his symptoms worsened, the patient experienced difficulty in walking. Neurological examination revealed abnormal sensation in both hands and lower limbs, tactile and hyperalgesia, total sensory insensitivity below the Th6 level, and a positive Romberg test. Blood tests showed elevated sIL-2R and ACE levels at 2,102 ‍U/ml and 23.7 ‍U/l, respectively. Antinuclear antibody, anti-SS-A antibody, MPO-ANCA, and PR3-ANCA tests were all negative. ELISA for AQP4-Ab was positive at 6.0 ‍U/ml. CSF analysis revealed an elevated cell count of 9/‍μl (5/‍μl mononuclear cells and 4/‍μl polymorphonuclear cells), protein concentration of 77.2 ‍mg/dl, and sIL-2R at 192 ‍U/ml, with a normal IgG index at 0.66. Whole-body CT revealed enlarged mediastinal and hilar lymph nodes (Fig. 2E, F), while ⁶⁷Ga scintigraphy demonstrated increased uptake in these areas with ‍mild accumulation in the myocardium. Transthoracic echocardiography revealed no ventricular septal thinning but indicated a decreased left ventricular ejection fraction of 49%. Spinal cord MRI revealed a high-signal lesion from the center of the Th6 spinal cord extending dorsally (Fig. 1G). There were no systemic symptoms such as fever or weight loss, nor laboratory abnormalities such as elevated lactate dehydrogenase levels. Furthermore, despite the presence of neurological symptoms, the lymphadenopathy was limited to the hilar region, which was considered atypical for lymphoma. Although a definitive patho­logical diagnosis was not established, the presence of systemic findings, such as enlarged hilar lymph nodes, severe atrioventricular block, and left ventricular systolic dysfunction, along with elevated serum sIL-2R and ACE levels, indicated systemic sarcoidosis. Because no other etiology was identified for the spinal cord lesion, a unified diagnosis of possible neuro­sarcoidosis was made. Furthermore, AQP4-Ab retesting by CBA yielded negative results, confirming that the initial ELISA test was a false positive. Following one course of IVMP at 1,000 ‍mg/day for 3 days, PSL at 30 ‍mg/day and MTX at 2 ‍mg/week were initiated. The patient’s symptoms gradually improved, and he regained the ability to walk. The patient was maintained on PSL at 5 ‍mg/day and MTX at 6 ‍mg/week without recurrence.

Discussion

In this study, we encountered three cases of neurosarcoidosis that initially tested positive for AQP4-Ab: two cases by ELISA and one case by CBA (Table 1).

Table 1 Summary of three cases.

case	diagnosis	age	sex	neurological symptoms	AQP4-Ab (ELISA)	AQP4-Ab (CBA)	systemic symptoms	sIL2-R (serum)	sIL2-R (CSF)	treatment	outcome
1	probable NS	66	M	bilateral optic perineuritis	negative	positive a)	enlargement of the mediastinal and hilar lymph nodes	1,699 ‍U/ml	153 ‍U/ml	IVMP, PSL, IFX b)	improvement
2	definite NS	66	M	myelitis	5.8 ‍U/ml	negative	enlargement of the mediastinal and hilar lymph nodes	1,176 ‍U/ml	121 ‍U/ml	IVMP, PSL, MTX	stable
3	possible NS	76	M	myelitis	6.9 ‍U/ml	netgative	enlargement of the mediastinal and hilar lymph nodes	2,102 ‍U/ml	192 ‍U/ml	IVMP, PSL, MTX	improvement

^a)Negative after treatment. ^b)Ravulizumab and inebilizumab were used before the diagnosis of neurosarcoidosis was made. Abbreviations: AQP4-Ab: aquaporin-4 antibody, ELISA: enzyme-linked immunosorbent assay, CBA: cell-based assay, sIL-2R: soluble interleukin-2 receptor, CSF: cerebrospinal fluid, NS: neurosarcoidosis, M: male, F: female, IVMP: intravenous methylprednisolone, PSL: prednisolone, IFX: infliximab, MTX: methotrexate.

Neurosarcoidosis and NMOSD can both present with optic nerve and spinal cord involvement, emphasizing the significance of differentiation. ELISA has a sensitivity of 63–83.3% and a false-positive rate of 0.5–1.3%⁵⁾⁸⁾⁹⁾. In contrast, CBA has a sensitivity of 76.7% and a false-positive rate of 0.1%⁵⁾¹⁰⁾, making it the superior and recommended diagnostic method. However, CBA requires specialized equipment and can be technically and economically challenging in some cases¹¹⁾.

In Case 1, AQP4-Ab was negative by ELISA but positive by CBA, which was performed for the first time. However, several clinical features were not consistent with a diagnosis of NMOSD. These included the presence of optic perineuritis, repeated relapses despite treatment with multiple biologic agents, including ravulizumab, and favorable response to corticosteroid therapy without residual neurological deficits after each recurrence. Furthermore, histopathological examination of mediastinal lymph node biopsy specimens revealed non-caseating epithelioid granulomas, supporting a diagnosis of probable neurosarcoidosis based on the international diagnostic criteria³⁾. Although there have been reports of coexisting NMOSD and sarcoidosis¹²⁾¹³⁾, the clinical presentation in this case was more consistent with neurosarcoidosis alone. Repeat testing for AQP4-Ab by CBA after treatment yielded a negative result. These findings raise the possibility that the initial AQP4-Ab result was truly positive but became negative after treatment. Majed et al.¹⁴⁾ reported that 103 (11%) of 933 patients who underwent repeated AQP4-Ab testing exhibited seronegative conversion during the disease course, with younger age and lower initial antibody titers associated with this transition. Although the antibody titer was not quantified in our case, it may have been below the detection threshold of ELISA, and treatment could have further reduced the titer, resulting in a negative CBA on retesting. On the other hand, although CBA is generally considered highly specific, false positives have been reported, particularly in patients with multiple sclerosis who received natalizumab¹⁵⁾. Therefore, a false-positive CBA result cannot be completely ruled out in this case. A repeat test using the CBA on the same specimens was considered to confirm the possibility of false-positive results; however, this could not be performed due to the unavailability of stored samples.

In Cases 2 and 3, AQP4-Ab was positive by ELISA but negative by CBA using pretreatment specimens, strongly suggesting false-‍positive ELISA results. The antibody titers in both cases were low (5.8 and 6.0 ‍U/ml, respectively), further supporting this ‍interpretation. False-positive AQP4-Ab results leading to misdiagnosis have been previously reported. For example, Suthiphosuwan et al.¹⁷⁾ described false-positive ELISA results in patients with spinal arteriovenous fistulas. However, to the best of our knowledge, no prior studies have reported false-positive AQP4-Ab results in neurosarcoidosis, either by ELISA or CBA. In a study by Flanagan et al.⁶⁾, AQP4-Ab testing in 19 patients with spinal sarcoidosis, who presented with longitudinally extensive spinal cord lesions requiring differentiation from NMOSD, yielded negative results in all cases. Therefore, in cases of suspected neurosarcoidosis with low-titer AQP4-Ab positivity by ELISA, the possibility of false-positive results should be considered. Confirmatory testing using CBA is recommended to avoid misdiagnosis.

The clinical characteristics of the cases we experienced are that Cases 2 and 3 elevated sIL-2R levels, and enlarged hilar lymph nodes at the time of the onset of neurological symptoms, in contrast, Case 1 exhibited these findings later in the disease course. All three cases demonstrated elevated CSF sIL-2R levels. Although sIL-2R levels in the CSF are nonspecific and may be elevated in other conditions, such as tuberculous meningitis and central nervous system lymphoma¹⁸⁾, Ikeguchi et al.¹⁹⁾ found that the median CSF sIL-2R levels in 30 cases of NMOSD was <54.5 ‍U/ml, with no observed increases. Additionally, it has been reported that CSF sIL-2R levels are elevated in more than 50% of patients with neurosarcoidosis, with a median value of 61.1 ‍U/ml (range: 0.9–6,970.2)¹⁸⁾. The possibility of neurosarcoidosis should be considered in patients with positive AQP4-Ab results and suspected NMOSD, particularly when elevated blood or CSF sIL-2R levels or hilar lymphadenopathy are observed.

Conflict of interest

The authors declare that there is no conflict of interest relevant to this article.

References

• 1)  Iannuzzi MC, Rybicki BA, Teirstein AS. Sarcoidosis. N Engl J Med 2007;357:2153-2165.
• 2)  Baughman RP, Teirstein AS, Judson MA, et al. Clinical characteristics of patients in a case control study of sarcoidosis. Am J Respir Crit Care Med 2001;164:1885-1889.
• 3)  Stern BJ, Royal W, Gelfand JM, et al. Definition and consensus diagnostic criteria for neurosarcoidosis: from the neuro­sarcoidosis consortium consensus group. JAMA Neurol 2018;75:1546-1553.
• 4)  Nozaki K, Scott TF, Sohn M, et al. Isolated neurosarcoidosis: case series in 2 sarcoidosis centers. The Neurologist 2012;18:373-377.
• 5)  Wingerchuk DM, Banwell B, Bennett JL, et al. International consensus diagnostic criteria for neuromyelitis optica spectrum disorders. Neurology 2015;85:177-189.
• 6)  Flanagan EP, Kaufmann TJ, Krecke KN, et al. Discriminating long myelitis of neuromyelitis optica from sarcoidosis. Ann Neurol 2016;79:437-447.
• 7)  Kidd DP, Burton BJ, Graham EM, et al. Optic neuropathy associated with systemic sarcoidosis. Neurol Neuroimmunol Neuroinflammation 2016;3:e270.
• 8)  Pittock SJ, Lennon VA, Bakshi N, et al. Seroprevalence of aquaporin-4-IgG in a northern California population repre­senta­tive cohort of multiple sclerosis. JAMA Neurol 2014;71:1433-1436.
• 9)  Tampoia M, Abbracciavento L, Barberio G, et al. N. A new M23-based ELISA assay for anti-aquaporin 4 autoantibodies: diagnostic accuracy and clinical correlation. Autoimmun Highlights 2019;10:5.
• 10)  Waters P, Reindl M, Saiz A, et al. Multicentre comparison of a diagnostic assay: aquaporin-4 antibodies in neuromyelitis optica. J Neurol Neurosurg Psychiatry 2016;87:1005-1015.
• 11)  Fu Y, Bi J, Yan Y, et al. Rapid Immunodot AQP4 assay for neuromyelitis optica spectrum disorder. JAMA Neurol 2023;80:1105-1112.
• 12)  Sawaya R, Radwan W. Sarcoidosis associated with neuromyelitis optica. J Clin Neurosci Off J Neurosurg Soc Australas 2013;20:1156-1158.
• 13)  Foster MA, Collorone S, Palace J, et al. Sarcoidosis and neuromyelitis optica in a patient with optic neuritis - a case report. Ann Clin Transl Neurol 2021;8:1760-1763.
• 14)  Majed M, Valencia Sanchez C, Bennett JL, et al. Alterations in aquaporin-4-IgG serostatus in 986 patients: a laboratory-based longitudinal analysis. Ann Neurol 2023;94:727-735.
• 15)  Cohen M, De Sèze J, Marignier R, et al. False positivity of anti aquaporin-4 antibodies in natalizumab-treated patients. Mult Scler Houndmills Basingstoke Engl 2016;22: 1231-1234.
• 16)  Williams JP, Abbatemarco JR, Galli JJ, et al. Aquaporin-4 autoantibody detection by ELISA: a retrospective character­ization of a commonly used assay. Mult Scler Int 2021;2021:8692328.
• 17)  Suthiphosuwan S, Oh J, Bharatha A. Clinical pitfall: false-positive aquaporin-4 IgG leading to misdiagnosis of neuromyelitis optica spectrum disorder in patient with spinal arteriovenous fistula. Spinal Cord Ser Cases 2017;3:17030.
• 18)  Otto C, Wengert O, Unterwalder N, et al. Analysis of soluble interleukin-2 receptor as CSF biomarker for neurosarcoidosis. Neurol Neuroimmunol Neuroinflammation 2020;7:e725.
• 19)  Ikeguchi R, Shimizu Y, Shimizu S, et al. CSF and clinical data are useful in differentiating CNS inflammatory demyelinating disease from CNS lymphoma. Mult Scler Houndmills Basingstoke Engl 2018;24:1212-1223.