Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
September 2009

Prediction of Neuromyelitis Optica Attack Severity by Quantitation of Complement-Mediated Injury to Aquaporin-4–Expressing Cells

Author Affiliations

Author Affiliations: Departments of Laboratory Medicine and Pathology (Drs Hinson, Apiwattanakul, Lennon, and Pittock and Mr Fryer), Neurology (Drs McKeon, Lennon, and Pittock), and Immunology (Dr Lennon), Mayo Clinic College of Medicine, Rochester, Minnesota.


Copyright 2009 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2009

Arch Neurol. 2009;66(9):1164-1167. doi:10.1001/archneurol.2009.188

Background  Recent reports support a pathogenic role in neuromyelitis optica (NMO) for the aquaporin-4 (AQP4)–specific autoantibody (NMO-IgG). Neuromyelitis optica is an inflammatory demyelinating central nervous system disease, usually relapsing, that causes variable degrees of attack-related disability. The NMO-IgG binds in vitro to the extracellular domain of AQP4, activates complement, and causes astrocyte lesioning.

Objective  To compare the prognostic utility of NMO-IgG titer and quantitative measures of complement-mediated injury to AQP4-expressing cells in NMO attacks.

Design, Setting, and Participants  A retrospective clinical-serological correlative study at Mayo Clinic's Neuroimmunology Laboratory was undertaken. Over an 18-month period, we identified NMO-IgG–seropositive patients in whom sufficient serum and adequate clinical information pertaining to NMO attacks (6 severe, 6 mild) were available to analyze clinical-serological correlations. Sera from 9 patients with multiple sclerosis and 9 healthy subjects (all NMO-IgG seronegative) served as controls. Complement activation was measured by quantifying the number of green fluorescent protein–AQP4–transfected HEK 293 cells permeable to the viability dye propidium iodide after exposure to patient serum and active complement.

Main Outcome Measures  Attack severity (mild or severe), percentage of AQP4-transfected cells lesioned, and NMO-IgG titer.

Results  The median percentage of AQP4-transfected cells lesioned by complement in the presence of serum from patients with NMO was 14% for patients with mild attacks and 54% for patients with severe attacks (P = .005). Median complement activation values for sera from healthy subjects and patients with multiple sclerosis were 8% and 12%, respectively. Patients with mild NMO attacks and patients with severe NMO attacks did not differ significantly with respect to NMO-IgG titer (P = .089).

Conclusions  A laboratory measure of complement-mediated cell injury may serve as a prognostic biomarker in NMO. Larger prospective studies are required to validate this observation.

Neuromyelitis optica (NMO) comprises a spectrum of relapsing autoimmune, inflammatory, demyelinating disorders of the central nervous system (CNS) and predominantly affects the optic nerves and spinal cord.1 The NMO-IgG autoantibody, detected by a tissue-based immunofluorescence assay, is 58% to 73% sensitive and 91% to 100% specific for clinically defined NMO.2,3 Its detection allows the distinction of NMO from multiple sclerosis (MS) and other idiopathic inflammatory CNS disorders. Patients who are NMO-IgG seropositive generally have a severe course, with rapid accrual of attack-related disability (typically blindness and paraplegia).1 Recovery is complete or nearly complete in milder NMO cases, even in the setting of frequent attacks.

There is growing evidence that NMO-IgG is a pathophysiological effector in NMO through its binding to the astrocytic transmembrane protein aquaporin-4 (AQP4), which is the predominant water channel in the CNS.4,5 A pathogenic role for NMO-IgG has been demonstrated in vitro1,58 and is strongly implied by immunohistochemical findings in patients' lesional CNS tissues. The NMO-IgG binds to the extracellular domain of the AQP4 water channel and initiates complement lesioning of astrocytes.5,6,8 In the absence of complement, NMO-IgG downregulates astrocyte surface expression of both AQP4 and the major CNS glutamate transporter, excitatory amino acid transporter 2, which is coupled to AQP4.6 Perturbation of water and glutamate homeostasis likely contributes to the pathogenesis of NMO. A loss of AQP4 from polarized end-feet of astrocytes where they contact endothelial cells of the blood-brain barrier could adversely alter osmotic regulation, blood-brain barrier function, granulocyte recruitment, and maintenance of the neural microenvironment.5,6,8

The NMO-IgG antibody titers may correlate with disease activity in NMO attacks in some patients followed longitudinally,9 and they may also be directly proportional to transverse myelitis lesion length and attack-nadir severity.10 To date, these findings have not been reproduced in large prospective studies.

Functional autoantibody assays are rare in neurological practice. A singular example is the muscle acetylcholine receptor modulating antibody assay used in evaluation of myasthenia gravis. It measures endocytosis and degradation of acetylcholine receptors initiated through cross-linking of cell surface acetylcholine receptors by bivalent antibodies. Acetylcholine receptor loss values of 90% and higher correlate significantly with weakness severity and the probability of an associated neoplasm.11,12

Assays to predict functional outcome in NMO might facilitate therapeutic decision making. Here we demonstrate the potential for a quantitative measure of complement activation by NMO-IgG to predict the clinical severity of an NMO attack.


Between January 1, 2007, and July 31, 2008, we retrospectively identified 12 NMO-IgG–seropositive patients (by tissue-based immunofluorescence assay) for whom detailed information was available. This information included clinical and radiological data, NMO-IgG titer, and timing of immunosuppressant therapies in relation to blood draw dates (Table). Attacks of NMO were mild in 6 patients (complete or near complete clinical recovery) and severe in 6 patients (blind or paraparetic with little or no clinical recovery). High-titered NMO-IgG–positive sera (pooled from a minimum of 50 patients) were used as positive controls; sera from 9 healthy subjects and 9 patients with classical MS served as disease specificity controls.

Image not available
Demographic, Clinical, and Serological Data for 12 NMO-IgG–Seropositive Patients

Complement activation was measured (blinded) by flow cytometric analysis of the percentage of green fluorescent protein–AQP4–transfected HEK 293 cells permeable to the viability dye propidium iodide after exposure to patient serum (20% serum for 30 minutes at room temperature) and active complement (20% complement for 45 minutes at 37°C; Low-Tox-H rabbit complement; Cedarlane Laboratories Ltd, Burlington, North Carolina) as previously described.5 Permeability to propidium iodide indicates a potentially lethal compromise in plasma membrane integrity. The total number of cells counted per condition was 10 000. We have previously demonstrated that untransfected HEK 293 cells (cells lacking AQP4 expression) are unaffected by exposure to NMO-IgG in the presence of active or inactive complement5 and that the increase in membrane permeability in AQP4-expressing cells exposed to NMO-IgG is dependent on active complement. To control for the requirement of active complement in this study, we included as both positive and negative controls pooled NMO-IgG–positive sera and normal sera as well as active or heat-inactivated complement.5 These experiments validated that NMO-IgG and active complement were both necessary for efficient membrane lesioning (data not shown).

The NMO-IgG titer (reciprocal of the final dilution scored positive in doubling serum dilutions by standardized immunofluorescence assay2) and degree of complement activation (percentage of propidium iodide–positive [lesioned] cells) were compared for patients with NMO who had mild and severe attack outcomes (Wilcoxon rank sum test). The relationship between the NMO-IgG titer and degree of complement activation was evaluated by calculating the Pearson correlation coefficient and by performing linear regression analysis, where degree of complement activation was the dependent variable. Statistical significance was defined as P < .05. All statistical analyses were performed using JMP version 7.0 statistical software (SAS Institute, Inc, Cary, North Carolina).


Patients with mild and severe attacks of NMO did not differ significantly with respect to age, sex, ethnicity, and therapy at the time of blood draw (Table). All patients with mild attacks of NMO and 4 of 6 patients with severe attacks of NMO had blood drawn within 4 weeks of attack onset. The median percentages of AQP4-transfected cells lesioned by complement in the presence of sera from patients with mild and severe NMO attacks were 14% (range, 8%-19%) and 54% (range, 21%-60%), respectively (P = .005) (Figure). The NMO-IgG titers for patients with mild NMO attacks and patients with severe NMO attacks did not differ significantly (P = .089). Median complement activation values for sera from healthy subjects and patients with MS were 8% (range, 8%-12%) and 12% (range, 9%-15%), respectively. The positive control NMO-IgG–positive serum pool had a mean value of 55% complement activation (data not shown).

Image not available

Serum samples obtained from patients with neuromyelitis optica (NMO) during a clinical relapse. Regardless of the NMO-IgG titer (P = .089), attack severity correlated significantly (P < .005) with the percentage of aquaporin-4–transfected cells lesioned by complement. Complement-activating activity of sera from patients with mild attacks did not differ significantly from that of healthy participants or patients with multiple sclerosis (MS). Dashes indicate individual patient values; dots, median values; and error bars, interquartile ranges.


Our preliminary analysis of data from testing sera of patients with NMO and control participants indicates a statistically significant association between NMO attack severity and measures of complement-mediated injury to AQP4-expressing cells. The results support our hypothesis that binding of NMO-IgG to AQP4 initiates an inflammatory cascade resulting in potentially irreversible neurological injury. Control sera did not significantly activate complement, and we did not find a statistically significant association between NMO-IgG titer and attack severity or between NMO-IgG titer and degree of complement activation.

Limitations of this study include the retrospective design and small numbers of patients tested, which may have precluded detection of a positive association between NMO-IgG titer and attack severity. Also, the technique used to measure the percentage of HEK 293 cells permeable to propidium iodide has not been validated as a quantitative measure of complement-mediated cell lesioning in a clinical service laboratory setting. Larger prospective studies correlating NMO-IgG characteristics (including antibody titers and quantitative functional effects) with clinical course or attack severity may provide prognostic serological profiles. For the individual patient, these profiles might identify novel therapeutic targets and direct therapeutic decision making.

No data are currently available to support the efficacy of a specific therapeutic modality for NMO.1 Current recommendations are based on small case numbers and anecdotal experience. Nevertheless, it has been established that disability in NMO is attack related and that a secondary progressive course is uncommon.13 This contrasts with MS, where the main risk factor for developing permanent disability is the onset of a secondary progressive course, apparently independent of superimposed relapses. Thus, successful attack prevention therapy, if initiated early, should have a greater effect on the natural history of NMO than on that of MS. Therapies to date are limited to immunosuppressive agents (prednisone, azathioprine, mycophenolate mofetil, rituximab, and mitoxantrone hydrochloride) for attack prevention and to intravenous methylprednisolone and plasmapheresis for treatment of acute attacks. Our preliminary data may support consideration of complement-inhibitory therapies in patients with severe attacks.

Back to top
Article Information

Correspondence: Sean J. Pittock, MD, Departments of Neurology and Laboratory Medicine and Pathology, Hilton 3-78D, Mayo Clinic, 200 First St SW, Rochester, MN 55905 (

Accepted for Publication: February 11, 2009.

Author Contributions:Study concept and design: Hinson, McKeon, Fryer, and Pittock. Acquisition of data: Hinson, McKeon, and Apiwattanakul. Analysis and interpretation of data: Hinson, McKeon, Lennon, and Pittock. Drafting of the manuscript: Hinson, McKeon, and Pittock. Critical revision of the manuscript for important intellectual content: Fryer, Apiwattanakul, Lennon, and Pittock. Obtained funding: Pittock. Administrative, technical, and material support: Hinson, Fryer, Lennon, and Pittock. Study supervision: Pittock.

Financial Disclosure: Drs Hinson, Lennon, and Pittock are named inventors on a filed patent relating to functional assays for detecting AQP4-IgG. In accordance with the Bayh-Dole Act of 1980 and Mayo Foundation policy, Dr Lennon stands to receive royalties for intellectual property related to the AQP4 autoantigen. This intellectual property is licensed to a commercial entity for development of a simple antigen-specific assay to be made available worldwide for patient care. The test will not be exclusive to Mayo Clinic. To date, the authors have received a total of less than $10 000 in royalties. Mayo Clinic offers the test as an indirect immunofluorescence assay to aid the diagnosis of NMO, but the authors do not benefit personally from the performance of the test. The other authors have no conflicting financial interests.

Funding/Support: This work was supported by the Guthy-Jackson Charitable Foundation.

Additional Contributions: Sarah M. Jenkins, MS, Division of Biomedical Statistics and Informatics, Mayo Clinic, Rochester, Minnesota, reviewed our statistical methods, results, and interpretations. We thank all members of the Mayo Clinic NMO Consortium for stimulating discussions.

Wingerchuk  DMLennon  VALucchinetti  CFPittock  SJWeinshenker  BG The spectrum of neuromyelitis optica. Lancet Neurol 2007;6 (9) 805- 815
Lennon  VAWingerchuk  DMKryzer  TJ  et al.  A serum autoantibody marker of neuromyelitis optica: distinction from multiple sclerosis. Lancet 2004;364 (9451) 2106- 2112
Waters  PJarius  SLittleton  E  et al.  Aquaporin-4 antibodies in neuromyelitis optica and longitudinally extensive transverse myelitis. Arch Neurol 2008;65 (7) 913- 919
Lennon  VAKryzer  TJPittock  SJVerkman  ASHinson  SR IgG marker of optic-spinal multiple sclerosis binds to the aquaporin-4 water channel. J Exp Med 2005;202 (4) 473- 477
Hinson  SRPittock  SJLucchinetti  CF  et al.  Pathogenic potential of IgG binding to water channel extracellular domain in neuromyelitis optica. Neurology 2007;69 (24) 2221- 2231
Hinson  SRRoemer  SFLucchinetti  CF  et al.  Aquaporin-4-binding autoantibodies in patients with neuromyelitis optica impair glutamate transport by down-regulating EAAT2. J Exp Med 2008;205 (11) 2473- 2481
Roemer  SFParisi  JELennon  VA  et al.  Pattern-specific loss of aquaporin-4 immunoreactivity distinguishes neuromyelitis optica from multiple sclerosis. Brain 2007;130 (pt 5) 1194- 1205
Vincent  TSaikali  PCayrol  R  et al.  Functional consequences of neuromyelitis optica-IgG astrocyte interactions on blood-brain barrier permeability and granulocyte recruitment. J Immunol 2008;181 (8) 5730- 5737
Jarius  SAboul-Enein  FWaters  P  et al.  Antibody to aquaporin-4 in the long-term course of neuromyelitis optica. Brain 2008;131 (pt 11) 3072- 3080
Takahashi  TFujihara  KNakashima  I  et al.  Anti-aquaporin-4 antibody is involved in the pathogenesis of NMO: a study on antibody titre. Brain 2007;130 (pt 5) 1235- 1243
Howard  FM  JrLennon  VAFinley  JMatsumoto  JElveback  LR Clinical correlations of antibodies that bind, block, or modulate human acetylcholine receptors in myasthenia gravis. Ann N Y Acad Sci 1987;505526- 538
Lennon  VA Serologic profile of myasthenia gravis and distinction from the Lambert-Eaton myasthenic syndrome. Neurology 1997;48 ((suppl 5)) S23- S27Article
Wingerchuk  DMPittock  SJLucchinetti  CFLennon  VAWeinshenker  BG A secondary progressive clinical course is uncommon in neuromyelitis optica. Neurology 2007;68 (8) 603- 605