Case
An 18-month-old Japanese boy without significant past history was admitted to our hospital with a 1-week history of limb tremor, opsoclonus and ataxia. He was unable to sit or walk without support. He had no history of vaccinations or infections within the preceding month. Computed tomography revealed a localized anterior sacral tumor and no lesions of the central nervous system. The anterior sacral tumor showed avid uptake of 123I-metaiodobenzylguanidine (123I-MIBG). Blood testing revealed none of the autoantibodies reported in association with paraneoplastic neurological syndromes such as anti-Hu, anti-Ri, anti-Yo, anti-amphiphysin, anti-PNMA2, anti-CV2, anti-recoverin, anti-SOX1, anti-titin, anti-zic4, anti-GAD65 and anti-Tr.[5] CSF analysis revealed an increased number and ratio of CD20-positive B cells, at 0.66 cells/µl and 12%, respectively (reference B-cell ratio in CSF: 0.00–0.03%[6]). In addition, oligoclonal bands and anti-GluRδ2 antibodies were detected in CSF.[2,3] The anterior sacral tumor was resected completely and diagnosed pathologically as neuroblastoma without MYCNamplification, classified as very low risk group in the International Neuroblastoma Risk Group Classification. In addition, detailed pathological analysis revealed multiple lymphoid follicles containing abundant CD20-positive B cells in the neuroblastoma tissue (Fig. 1). The increased CD20-positive B cells, oligoclonal bands and anti-GluRδ2 antibodies in CSF, and multiple lymphoid follicles containing abundant CD20-positive B cells in neuroblastoma tissue led us to consider the possibility that humoral immunity may contribute to OMS.
OMS has been reported to persist after complete resection of neuroblastoma and to require immunosuppressive therapy after tumor resection. We started administration of monthly dexamethasone (DEX) (20 mg/m2/day for 3 days) pulses and IVIg (1 g/kg/day) according to a previous report.[7] To assess the effectiveness of treatment, we monitored the severity of OMS symptoms using “OMS Rating Scales”, comprising evaluations of six symptoms: stance, gait, arm function, opsoclonus, mood/behavior and speech (Supplementary Table S1).[8] Simultaneously, we monitored CD20-positive B cells and oligoclonal bands in CSF over time, as factors reported to reflect the disease activity of OMS. After starting immunosuppressive therapy, limb tremor, ataxia and opsoclonus gradually remitted. After three cycles, the patient was able to walk again without support. OMS Rating Scales score also improved to 2, from 8 at the first visit. Although oligoclonal bands in CSF had been detected persistently, the B-cell ratio in CSF declined in parallel with symptom improvement (Fig. 2).
Since significant improvements by DEX pulse and IVIg therapy and increased CD20-positive B cells in CSF and multiple lymphoid follicles with abundant CD20-positive B cells in tumor tissue were observed, we considered the existence of identical B-cell clones contributed to the pathogenesis of OMS in both CSF and tumor tissues. We therefore performed BCR repertoire analysis of B-cells in both CSF and tumor tissue using unbiased amplification of BCR genes followed by next-generation sequencing (Repertoire Genesis Inc., Osaka, Japan) as previously described.[9] Consequently, the existence of identical B-cell clone was confirmed in both CSF and tumor tissue (Supplemental Table S2). Intriguingly, the identical B-cell clone was the dominant clone in both CSF and tumor tissue. Moreover, various diversity indices including the Shannon Weaver index, inverse Simpson index, Pielou’s evenness and Diversity Evenness 50 were applied to evaluate the diversity and clonality of B-cells in both CSF and tumor tissue. As a result, every diversity index confirmed that B-cells in CSF showed low diversity and B-cells in tumor tissue showed high diversity (Supplemental Table S3). These data suggested that low-diversity B-cell clones migrating into the central nervous system from tumor tissue may express autoantibodies such as anti-GluRδ2 antibodies, leading to the pathogenesis of OMS.