Screening for onconeural antibodies in neuromyelitis optica spectrum disorders

Background Some so-called “non-classical” paraneoplastic neurological syndromes (PNS), namely optic neuritis and myelitis, clinically overlap with neuromyelitis optica spectrum disorders (NMOSD), and conversely, in cancer-associated NMOSD, a paraneoplastic etiology has been suggested in rare cases. Therefore, we retrospectively investigated the prevalence of onconeural antibodies, which are highly predictive for a paraneoplastic etiology, and the prevalence of malignancies in NMOSD patients. Methods We retrospectively screened 23 consecutive patients from our clinic with NMOSD (13 were anti-aquaporin-4 [AQP4] antibody positive, 10 were AQP4 negative) for onconeural antibodies using an immunoblot. Results All patients were negative for a broad spectrum of antibodies targeting intracellular onconeural antigens (Hu, Yo, Ri, CV2/CRMP5, Ma1, Ma2, Zic4, SOX1, Tr, and amphiphysin). Notably, only two patients had a malignancy. However, neoplastic entities (astrocytic brain tumor and acute myeloid leukemia) were not typical for PNS. Conclusions Our data suggest that there is no need to routinely screen anti-AQP4 antibody positive NMOSD patients with a typical presentation for onconeural antibodies. Furthermore, absence of these antibodies in NMOSD, which is typically non-paraneoplastic, confirms their high specificity for PNS.


Background
Neuromyelitis optica (NMO) is a rare, immunemediated, demyelinating disorder of the central nervous system (CNS), typically presenting with relapsing optic neuritis (ON) and/or ≥ three vertebral segment longitudinally extensive transverse myelitis (LETM) [1,2]. Pathogenetic antibodies targeting the water channel protein aquaporin-4 (AQP4) are found in the majority of patients with NMO [3]. Since their discovery, the spectrum of clinical manifestations within the CNS associated with AQP4 antibodies has expanded [4]. Therefore, diagnostic criteria have recently been revised, introducing the term "neuromyelitis optica spectrum disorders (NMOSD)" [5]. According to these revised criteria, an NMOSD diagnosis can also be established in absence of anti-AQP4 antibodies. For simplicity, in the following, the term "NMOSD" is consistently used for both NMO and NMOSD.
Regarding a previously suggested paraneoplastic etiology in rare cases, we retrospectively investigated the prevalence of onconeural antibodies and malignancies in NMOSD patients.

Methods
Consecutive patients were identified by an electronic database search. Based on clinical records, NMOSD diagnosis was verified according to recently revised criteria [5]. This approach identified 35 patients with NMOSD who were treated in our clinic (Department of Neurology and Neurophysiology, Medical Center-University of Freiburg, Germany) between 2003 and 2015. Stored serum samples kept at -80°C from 25 therapy naïve patients were available for analysis. Of these patients, two declined analysis. Finally, 23 patients entered the study. Demographic and clinical data, including anti-AQP4 antibody status, were obtained from patients' records.
Dichotomized variables are presented using numbers and percentages; continuous variables are presented using means or medians, range, and standard deviation (SD). The local ethics committee approved the study, and all patients gave written informed consent to the study protocol. Table 1 summarizes clinical data of 23 patients fulfilling revised criteria for NMOSD diagnosis and entering the study. Mean age was 44 years (range 19-75, SD 17.2) at disease manifestation, and 49 years (range 20-75, SD 15.8) at diagnosis. Eighteen (78.3%) were female, and 13 (56.5%) were anti-AQP4 antibody positive. Two patients ( Table 1: patients #5 and #15) had a malignoma: one had an anaplastic astrocytoma that occurred 7 years after NMOSD manifestation and that progressed to secondary glioblastoma; the other had acute myeloid leukemia (AML) that was treated with stem cell transplantation 4 years before the NMOSD manifestation. Follow-up information was available in all patients with a median duration of 5.0 years (range 0.5-10.0 years, SD 2.7). Remarkably, none had antibodies targeting intracellular onconeural antigens (Hu, Yo, Ri, CV2/CRMP5, Ma1, Ma2, Zic4, SOX1, Tr, and amphiphysin).
Limitations of our study were the retrospective design and therefore patients were not systematically screened for occult malignomas. Furthermore, the case number was limited, since serum was available for only 25 of 35 patients (71.4%) previously identified by an electronic database search for those with an NMOSD diagnosis.

Conclusions
According to our data, the routine screening for onconeural antibodies in NMOSD patients is not mandatory. However, clinicians should pay particular attention in anti-AQP4 negative patients, in patients with a known malignancy or cancer risk factors (e.g. smoking), and/or if clinical presentation is atypical, since paraneoplastic myelitis and/or ON in association with anti-CV2/ CRMP5, -Hu or -amphiphysin antibodies might clinically mimic NMOSD [7][8][9][10]. Finally, the absence of onconeural antibodies in a typically non-paraneoplastic disorder corresponds to their high specificity for PNS [6]. Finally, larger retrospective trials are necessary to verify these results and to determine the proportion of anti-AQP4 negative NMOSD patients with onconeural antibodies.