Tuesday, August 16, 2016

Neuroblastoma and multiple antibody-associated syndromes

Amini A, Lang B, Heaney D, Irani SR. Multiple sequential antibody-associated
syndromes with a recurrent mutated neuroblastoma. Neurology. 2016 Aug

A 5-year-old girl was diagnosed with opsoclonus-myoclonus syndrome (OMS) and a para-aortic neuroblastoma (NB). Urinary catecholamines were elevated; NB biopsy confirmed no MYCN gene amplification. Serum showed Hu antibodies without hippocampal neuron surface antibodies (NSAbs) and, of note, serum inhibited proliferation of cultured NB cells. After partial resection and chemotherapy, she progressed well in mainstream education. 
A decade later, she developed weight loss and abdominal pain followed by sudden right-sided sensorineural hearing loss with a gradually progressive syndrome of ataxia and opsoclonus, which responded transiently to corticosteroids. Subsequently, a generalized seizure, ataxia, multidirectional diplopia, and left arm sensorimotor disturbance prompted evaluation. This revealed bilateral medial temporal and mild right precentral gyrus high signal on MRI , moderately diffuse slowing on EEG, and serum Hu antibodies. Serum antibodies to GAD, Yo, or Ri, the VGKC complex, the NMDA, GABAA, GABAB, glycine [alpha], and AMPA receptors, tumor markers, and urinary catecholamines were not detected. Abdominal magnetic resonance and metaiodobenzylguanidine scintigraphy suggested NB recurrence. On this occasion, biopsy showed MYCN gene amplification and chromosome 1p deletion.

IV methylprednisolone (IVMP) plus NB debulking and radiotherapy allowed her to return to college feeling the best she had in 3 years, with resolution of signs and with normal neuroimaging.
One year later, she developed vomiting, dysphagia, and abdominal bloating. Serial imaging confirmed esophageal, colonic, and small bowel dilation with stable residual tumor appearances. This paraneoplastic chronic intestinal pseudo-obstruction (CIPO) remained refractory to medical therapies. Hu antibodies were found in serum and CSF and were the only reactivity observed on myenteric plexus staining. Alpha-3 acetylcholine ganglionic receptor antibodies were absent.

At age 22, worsening panintestinal dysmotility, cognition, and recurrent opsoclonus culminated in another seizure. Investigations showed new frontotemporal slow-wave activity, restricted CSF oligoclonal bands, and now, serum hippocampal NSAbs. IV methylprednisolone reduced frontotemporal activity and CSF Hu antibody titers, with transient symptom amelioration. Plasmapheresis achieved more sustained gastrointestinal and neurologic remission, permitting university studies continuation.

First, there is the striking co-occurrence of autoimmune syndromes associated with Hu antibodies and NB. These are likely to be autoantibody-mediated and are often paraneoplastic. Adult paraneoplastic OMS and limbic encephalitis (LE) with Hu antibodies are associated with breast cancer and small cell lung cancer. Usually, they show limited responses to immunotherapy.  By contrast, pediatric OMS typically associates with an NB of favorable prognosis and cytogenetics (absence of MYCN amplification) but rarely the Hu antibodies detected in our patient.  Pediatric Hu antibody–associated LE is usually nonparaneoplastic, whereas paraneoplastic Hu antibody–associated CIPO is usually reported with small cell lung cancer, with only rare reports in childhood NB. Paraneoplastic CIPO is important to differentiate from oncologic iatrogenic complications given its potential response to immunotherapy.  Abrupt-onset sensorineural hearing loss is usually an autoimmune steroid-responsive disorder, and has also been associated with Hu antibodies.  OMS itself has only rarely been associated with NSAbs.

Second, cytogenetic and biochemical characteristics of our patient's NB altered on recurrence. An alteration in the tumor biology, such as the de novo observed MYCN amplification and 1p deletion, may have broken immunologic tolerance and generated antibodies against NB-expressed antigens, also expressed in the hippocampi (causing LE) and the myenteric plexus (producing CIPO). This alteration may also be linked to the cessation of catecholamine secretion. Furthermore, a chromosome 1p deletion in childhood NB would usually signify poor prognosis, but since its partial resection, our patient has already survived 7 years and completed higher education. Indeed, our patient's serum inhibited proliferation of NB cells in culture. Collectively, these findings strengthen the established concept that an autoimmune response against the tumor may suppress its growth, as hypothesized in Lambert-Eaton myasthenic syndrome and classic anti-Hu antibody syndromes.  Also, maybe this ongoing antitumor response can expose neoantigens and lead to epitope spread. Ideally, this should be examined longitudinally using several serum and CSF samples but, unfortunately, these were not available from our patient.

Third, to our knowledge, this is the longest lag to the second presentation of NB with OMS and urges ongoing tumor vigilance during patient follow-up. Delayed recurrence of OMS after 9 years, and Hu antibody–associated LE 6 years after pediatric NB with OMS have been reported, with incomplete cytogenetic data.

Our case should prompt research correlating tumor cytogenetics and clinical progression in recurrent or spontaneously remitting tumors as a method to elucidate mechanisms linking antitumor and nervous tissue autoimmunity, tolerance mechanisms, and in ultimately identifying predictive signatures to guide therapy.


  1. Dale RC, Lim M. The origins and progression of CNS autoimmunity: Nature, nurture, and tumor. Neurology. 2016 Aug 9;87(6):560-1.

    Neuroblastoma is the most common extracranial solid cancer of childhood, accounting for 15% of cancer deaths in children. A small proportion (up to 4%) of children with neuroblastoma develop opsoclonus-myoclonus ataxia syndrome (OMAS), a paraneoplastic autoimmune CNS disorder. Of note, neuroblastoma associated with OMAS is frequently of low stage, incurring a better prognosis, suggesting that this tumor-triggered autoimmunity may also have a protective role.

    In this issue of Neurology, Amini et al. describe an intriguing and complex case that provides further glimpses at this important facet of CNS autoimmunity. The case in question initially presented at age 5 with OMAS, a para-aortic neuroblastoma, and anti-Hu antibodies; the tumor was successfully treated.

    Ten years later, the patient presented with a complex autoimmune CNS syndrome, with limbic encephalitis features plus opsoclonus and sensorineural deafness. The tumor had recurred and mutated, now with MYCN gene amplification (having previously been negative) and chromosome 1p deletion. Further tumor therapy plus IV methylprednisolone induced a surprisingly good recovery from the neurologic illness, allowing return to full time academic education. The patient developed a further suspected autoimmune complication 1 year later (chronic intestinal pseudo-obstruction), and then had a recurrence of the CNS syndrome, which responded to corticosteroids and plasma exchange, resulting in a return to normal functioning.(continued)

  2. (continued)In this patient, the breadth of evolving, overlapping, and relapsing nervous system features of suspected autoimmune origin, OMAS, limbic encephalitis, reversible sensorineural deafness, and pandysmotility syndrome (brain, ear, autonomic gut involvement) is highly unusual, although the co-occurrence of autoimmune nervous system syndromes is increasingly recognized, such as myasthenia gravis plus neuromyelitis optica and anti-NMDA receptor encephalitis plus CNS demyelination syndromes.

    The most interesting aspect of this case is the mutating neuroblastoma, which alongside other potential, yet unidentified, host factors is likely to have generated the diversity of autoimmune syndromes in this patient. Neural crest tumors make intellectually appealing triggers of paraneoplastic CNS disease, given the “neuronal-like” features of the tumor. Although there is currently little evidence of a “cell surface antibody” in the majority of patients with OMAS, the consistent immunotherapy responsiveness and histopathologic evidence of inflammatory infiltrate observed in neuroblastoma tumors associated with OMAS strongly support an autoimmune response both toward the tumor and also against the brain. Furthermore, this immunotherapeutic tumor effect has been successfully used in treating neuroblastoma with antibodies or, more recently, engineered T cells against highly expressed tumor antigen such as disialoganglioside (GD2).

    The impressive reversal of symptoms with only modest immune suppression at each relapse in this case is more typical of that seen in cell surface antibody–associated syndromes rather than classic paraneoplastic syndromes associated with onconeuronal antibodies (such as anti-Hu antibodies). It is conceivable, even likely, that this patient had a reversible immunologic process (humoral or cellular) throughout the disease course that was not detectable by current standard measures or antibody assays.
    The tumor mutation at the time of relapse in the patient reported by Amini et al. suggests that alteration of the cellular nature of the tumor is likely to have “reactivated” the paraneoplastic autoimmune CNS syndrome, by means of exposure of the host to “cryptic antigens” causing resultant loss of prior established immune tolerance, as hypothesized by the authors. This observation augments our understanding of the origins of autoimmune nervous system disease in a similar way to the recent recognition that the destructive pathology of herpes simplex virus encephalitis can provoke a secondary autoimmune response associated with anti-NMDA receptor antibodies or other cell surface antibodies,10 with the putative unifying process being the “release of previously cryptic neuronal antigens in an inflammatory milieu.”

    The careful longitudinal serum antibody analysis at presentation and through the relapsing phases of the disease in this case provides further insights into the natural history and evolution of the autoimmune process. Only anti-Hu antibodies were present throughout the clinical course, which are unusual in pediatric OMAS, and only rarely reported in children with inflammatory brain disease. Testing for a panel of cell surface antibodies was negative, and initial screening for a cell surface antibody using cultured live hippocampal neurons was negative at disease onset, although became positive in the relapsing stages of the illness over a decade after presentation. The fact that the patient only developed a potentially pathogenic cell surface antibody during the relapsing phase of disease further emphasizes the evolving nature of disease in this patient, hypothetically due to “epitope spreading” of the autoimmune response, potentially triggered by the mutating tumor. What is also intriguing is that during the early stages of disease, despite having no immunoglobulin binding to live hippocampal neurons, the patient's serum inhibited proliferation of neuroblastoma cells, suggesting the serum contained immunoglobulin or other immune or drug factors that had direct effects on neuronal-like cells.