The role of anti-aquaporin 4 antibody in the conversion of acute brainstem syndrome to neuromyelitis optica

Background Acute brainstem syndrome (ABS) may herald multiple sclerosis (MS), neuromyelitis optica (NMO), or occur as an isolated syndrome. The aquaporin 4 (AQP4)-specific serum autoantibody, NMO-IgG, is a biomarker for NMO. However, the role of anti-AQP4 antibody in the conversion of ABS to NMO is unclear. Methods Thirty-one patients with first-event ABS were divided into two groups according to the presence of anti-AQP4 antibodies, their clinical features and outcomes were retrospectively analyzed. Results Fourteen of 31 patients (45.16 %) were seropositive for NMO-IgG. The 71.43 % of anti-AQP4 (+) ABS patients converted to NMO, while only 11.76 % of anti-AQP4 (-) ABS patients progressed to NMO. Anti-AQP4 (+) ABS patients demonstrated a higher IgG index (0.68 ± 0.43 vs 0.42 ± 0.13, p < 0.01) and Kurtzke Expanded Disability Status Scale (4.64 ± 0.93 vs 2.56 ± 0.81, p < 0.01) than anti-AQP4 (-) ABS patients. Area postrema clinical brainstem symptoms occurred more frequently in anti-AQP4 (+) ABS patients than those in anti-AQP4 (-) ABS patients (71.43 % vs 17.65 %, p = 0.004). In examination of magnetic resonance imaging (MRI), the 78.57 % of anti-AQP4 (+) ABS patients had medulla-predominant involvements in the sagittal view and dorsal-predominant involvements in the axial view. Conclusions ABS represents an inaugural or limited form of NMO in a high proportion of anti-AQP4 (+) patients.

ABS is recognized more frequently in patients with NMO spectrum disorder (NMOSD), most brainstem symptoms exist even before the diagnosis of NMOSD [10][11][12]17]. In diagnostic criteria of NMOSD in 2015, ABS is considered as one of the core clinical characteristics in the diagnosis of NMOSD [1]. In this study, the predictive value of anti-AQP4 antibody for relapse or later development was tested after the first sign of ABS.

Study population
Thirty-one patients with first-event ABS admitted to the Third Affiliated Hospital of Sun Yat-sen University in Guangzhou, China from January 2009 to September 2011 were retrospectively analyzed. All the patients involved in this study fulfilled the following inclusive criteria: 1) test for NMO-IgG; 2) single clinical episode of ABS associated with relevant brainstem MRI lesions 3) no other neurologic signs or symptoms which suggested the diagnosis of MS or NMO before NMO-IgG testing. Exclusion criteria included previous or concomitant systemic autoimmune diseases, metabolic etiology, vascular disorders and infections. All the patients were negative for HIV antibody. The subjects provided written informed consent. This study was conducted according to the principles expressed in the Declaration of Helsinki and approved by the institution's ethics committee. Lumbar puncture was also performed with informed consent.

Data collection
Clinical features and outcomes including gender, age at onset, duration, relapse times, annualized recurrence rate, clinical manifestations, and magnetic resonance imaging (MRI) findings were recorded in details. Analysis of cerebrospinal fluid (CSF), serum anti-AQP4 test and examinations of magnetic resonance imaging (MRI) were performed within 2 weeks after attack of brainstem symptoms and before treatment. The diagnosis of NMO, NMOSD and MS was based on Wingerchuk's criteria in 2006 [18], International consensus diagnostic criteria for NMOSD in 2015 [1] and McDonald's criteria in 2010 [19] respectively. These symptoms, such as area postrema clinical syndrome (including intractable hiccups, nausea and vomiting), diplopia, and bulbar dysfunctions, were regarded as the manifestation of brainstem symptoms. A relapse of ABS was defined as definite brainstem symptoms of neurological dysfunction that lasted more than 24 h, together with relevant brainstem lesions after ruling out infective agents. Symptoms occurring within 1 month after the initial symptoms of relapse were considered to be part of the same episode. The neurological disability of the patients was assessed using the Kurtzke Expanded Disability Status Scale (EDSS) [20]. A corticosteroid (1000 mg methyl prednisone, administered intravenously for five consecutive days) was prescribed in the acute stage and some patients received azathioprine in the remission stage when developed as NMO.
AQP4 and oligoclonal bands testing AQP4 antibodies were determined using a cell-based assay on an AQP4-transfected cell line from a commercial BIOCHIP kit (EUROIMMUN AG, Lübeck, Germany) according to the manufacturer's instructions. The CSF oligoclonal bands (OCB) detection method used in our laboratory was an isoelectric focusing technique combined with the avidin-biotinperoxidase complex method.

MRI scanning
Brain and spinal cord MRI scans were carried out for all patients using a GE 1.5 T MR scanner (General Electric, Milwaukee, WI, USA). The slice thickness of the axial scans was between 3 and 5 mm. Conventional MRI protocols were used: T1 with and without gadolinium enhancement (400/15.5 ms, TR/TE) and T2 (2500-3500/100 ms, TR/TE) in spinal cord MRI; and T1 with and without gadolinium enhancement (2128-2300/11.6-12.4 ms, TR/TE), T2 (4600-4640/97.8-102 ms, TR/TE), and fluid-attenuated inversion recovery (FLAIR) (8800/120 ms, TR/TE) in brain MRI. A cross-sectional evaluation was also performed on all MRI scans of the brain, and brainstem lesions were classified as either having a ventral pattern, a dorsal pattern. The spinal cord was segmented into cervical and thoracolumbar regions. LETM is a spinal cord lesion that extends over 3 or more vertebral segments. All MRI scans were carried out prior to use of corticosteroid, immunomodulatory or immunosuppressive treatment. An experienced neuroradiologist and a neurologist, both of whom were blinded to the diagnostic categorization and the patients' clinical features, each analyzed all of the MRI scans. The final assessments were made by consensus.

Statistical analysis
All statistical analyses were performed using Statistical Program for Social Sciences (SPSS) statistical software (version 16.0; Chicago, IL, USA). For each set of values, data were expressed as the means ± standard deviation (SD). Categorical variables were compared using Fisher's exact test. Non-categorical variables were compared using the Mann-Whitney U test. Survival was estimated according to the Kaplan-Meier method. The primary study endpoint was the time to NMO conversion, as indicated by the Wingerchuk's criteria [18]. The log-rank test was used to compare the survival analysis between anti-AQP4 (+) and anti-AQP4 (-) ABS patients. All pvalues were two-tailed, with values of < 0.05 considered significant. This study was an exploratory study so that no adjustment for multiple comparisons was made.

Results
The data of 352 patients in our database were reviewed between 2009 and 2011. A total of 31 patients who were enrolled in this study satisfied the diagnostic criteria: 14 anti-AQP4 (+) patients with ABS and 17 anti-AQP4 (-) patients with ABS. The details of the enrollment process can be seen in the flowchart (Fig. 1).

CSF examinations
CSF specimens were obtained from all patients at the acute stage (Table 1)  significantly higher than that in anti-AQP4 (-) patients (log rank 5.23, p = 0.012) (Fig. 4).

Discussion
NMO frequently begins with an acute or subacute episode of ON or myelitis. However, brainstem symptoms in NMO were not rare [8,[10][11][12]17]. According to the previous studies, brainstem symptoms have recently been described in NMO as well as in NMOSD [10,12,13,17,21]. In 17.05-30.61 % NMO patients, brainstem symptoms even act as the sole manifestation [14,17]. "Brainstem symptoms with anti-AQP4 antibodies" has also been recommended in the latest diagnostic criteria for NMOSD [1]. Various brainstem symptoms such as area postrema clinical brainstem symptoms (including intractable hiccups, nausea and vomiting), diplopia, and bulbar dysfunctions attributed to lesions in the dorsal region of the medullar and the pons surrounding the fourth ventricle could be the first manifestation of symptoms [9,10,12,14]. Our results demonstrated that area  postrema clinical brainstem symptoms in anti-AQP4 (+) ABS patients were more frequent than those in anti-AQP4 (-) ABS patients. Intractable hiccups, nausea, and vomiting have been reported as unique symptoms in NMO due to the involvement of the pericanal region in the medulla oblongata, which included the areas postrema, nucleus tractus solitaries where the putative hiccup and vomiting centres are located [13,14]. In our study, intractable hiccups, nausea, and vomiting occur in 71.43 % anti-AQP4 (+) ABS patients, which are more frequent than in anti-AQP4(-) ABS patients. Previous study also confirmed that intractable hiccups, nausea, and vomiting are clinical markers for the early phase of an exacerbation in NMO [22]. In addition, anti-AQP4 (+) ABS patients are prone to have recurrent brainstem symptoms before attacks of ON or myelitis. In addition to clinical features, the distribution of brainstem lesions of anti-AQP4 (+) ABS patients also is different from anti-AQP4 (-) ABS patients. Although the midbrain and pons in the sagittal view are involved both in anti-AQP4 (+) ABS patients and anti-AQP4 (-) ABS patients, the medulla segments are more frequently involved in anti-AQP4 (+) ABS patients. In the axial view, anti-AQP4 (+) ABS patients present dorsal-predominant involvement, which is consistent with our previous report [23]. In AQP4 (+) ABS patients, majority of the   [11,17]. NMO-IgG has been proved to be highly specific and moderately sensitive to NMO. The prevalence of anti-AQP4 antibodies in NMO patients ranged from 50-60 % [3,[24][25][26]. However, NMO-IgG is not restricted to these patients fulfilling all criteria for a definite diagnosis of NMO [27][28][29]. This antibody is also identified in partial syndromes, the frequency of anti-AQP4 in acute partial transverse myelitis is 4.5 % [30], while 26.9 % in recurrent ATM [31], increasing to 37.9-60 % in LETM [27,31,32] and 20 % in recurrent ON [33], 5.8 % in acute monosymptomatic ON [34]. Our study provided that the frequency of anti-AQP4 in ABS is 45.16 %. Currently, sensitive marker(s) to predict the conversion of ABS to NMO are still absent. Our results found that 100 % of the anti-AQP4 (+) ABS patients experience clinical relapse, and 71.43 % of anti-AQP4 (+) ABS patients convert to NMO in the following 3 years, while only 11.76 % of the anti-AQP4 (-) ABS patients convert to NMO, and the cumulative NMO conversion probability is significantly higher in anti-AQP4 (+) ABS patients than that in anti-AQP4 (-) ABS patients. These results are similar to those patients with ON or LETM as the initial presentation, which also demonstrated a rapid NMO conversion in the first 2 years after the initial episode [32,35]. The results suggested that anti-AQP4 antibodies may be useful for diagnosis and prognosis of patients who presented with isolated ABS. Moreover, 29.42 % of anti-AQP4 (+) ABS patients experience recurrent brainstem symptoms before attack of ON or myelitis, which suggested that anti-AQP4 antibody associated recurrent brainstem symptoms should be paid more attention. The anti-AQP4 (+) ABS patients are at high risk for development of severe disability.
Several studies have also identified a new antigenic target, myelin oligodendrocytic glycoprotein (MOG), as being of interest in seronegative AQP4 patients [36]. Compared with AQP4-positivity patients, serum MOG antibodies do have a distinct clinical phenotype from AQP4-positivity NMO that is characterized by fewer relapses, a better clinical outcome, and a wider spectrum of MRI features, which may suggest a sort of disease distinct from NMO or MS [37][38][39]. This antibody is found in around 20 % of AQP4-seronegative patients. MOG testing was not performed in this study, therefore we can't exclude the bias of MOG-seropositive patients in our AQP4 negative cohort. However, brainstem lesions, which are a hallmark of AQP4-seropositive NMO/ NMOSD, occur less frequently in MOG-seropositive patients [38,40]. Moreover, the primary endpoint of our study is the conversion of typical NMO. Therefore, this would make little impact on our results.
Therapy of NMO should be initiated early. Azathioprine and rituximab are suggested as first-line treatments [41]. In patients with NMOSD, IFN-β treatment [42], glatiramer acetate [43], natalizumab [44] and fingolimod (FTY720) [45] are ineffective for preventing relapses or fail to control disease activity. In our study, no patients received the above treatments, which might have treatment effects on accumulative NMO conversion probability. Therefore, the risk of conversion to NMO in anti-AQP4 (+) ABS patients seems higher in the first 2 years based on this study. We recommend that all patients with ABS make test for NMO-IgG and that seropositive patients receive immunosuppressive treatments which are effective in reducing NMO relapse frequency, such as azathioprine or rituximab.

Conclusions
ABS represents an inaugural or limited form of NMO in a high proportion of anti-AQP4 (+) patients. Availability of data and materials All the data supporting our findings is contained within the manuscript.