Anti–Ri antibodies and limbic encephalitis in a patient with carcinoid tumour of the lung

Abstract

Sirs: The anti-Ri syndrome is defined as the coexistence of typical paraneoplastic neurological symptoms and anti-Ri antibodies. These antibodies typically occur in patients with a subacute neurological disorder usually in association with carcinoma of breast or lung [3–5]. The associated tumours seem to express ectopic neuro-onconal proteins, initiating an immune response cross reacting with neuronal nuclear tissue [6]. The clinical value of anti-Ri antibodies is their indicator function for a so far undetected tumour [7]. A 68-year-old man presented in August 2002, with slowly progressive personality changes and neuropsychological disorders. In February 1992 an endobronchial carcinoid tumour (pT2pN0) of the left upper lobe of the lung was resected in toto without evidence of metastasis on follow-up. Onset of neuropsychological symptoms was approximately July 2000. In August 2002 the patient scored 20 of 30 points in Folstein’s mini-mentalstate test. Detailed neuropsychological testing showed distinct apraxia, impaired short-term memory and deficits in complex actions and spatial-constructive exercises. Physical examination was otherwise unremarkable. Cranial magnetic resonance imaging (MRI) revealed modest fronto-temporal brain atrophy. Electroencephalogram (EEG) demonstrated slight general slowing and a focal slowing with continuous deltawaves fronto-temporal on the left and intermittent delta-waves temporal on the right side. Cerebrospinal fluid (CSF) showed a normal cell count, a slightly disturbed blood brain barrier (total protein of 48 mg/dL (< 45 mg/dL)), and negative oligoclonal bands. Examination excluded neurotropic viruses, bacteria or fungi in serum or CSF. Vasculitis screening and thyroid function were unobtrusive. Liver enzymes (GOT, GPT) were slightly increased while routine haematology and biochemistry were otherwise normal. Immunofluorescence analysis (IFA) with neuronal tissue revealed a high concentration of a serum antibody reactive with a nuclear antigen on cerebral and cerebellar sections, showing a typical pattern for anti-Ri antibody by Western blot (in house immunoblot with cerebellum extract and recombinant Ri-antigen). Serial dilutions revealed a specific antibody titre of 1:25600. In addition, the antibodies were identified as anti-Ri antibodies by a commercial immunofluorescence test, following the recommendations of the manufacturers (Euroimmun, Germany) and by a dot blot assay (onconeuronal antigens, Milenia Biotec, Germany). The serum was not reactive with tissue of peripheral nerve in IFA. Screening by CT and MRI revealed an asymptomatic metastatic invasion of multiple segments of the patient’s liver, lingula of the lung, and osteoplastic metastasis of several thoracal and lumbal vertebral bodies and of the sacrum. Bone scintigram was normal. Hydroxyindoleacetic acid was tenfold elevated in a 24 hour urine sample and neuron-specific enolase was elevated to levels of 17.7 μg/L (< 12.5 μg/L). Due to a negative somatostatin-receptor scintigram, treatment by somatostatin was not given. Biphosphonates were given intravenously every four weeks to prevent pathological bone fractures. 12 months after discharge the patient reported (by telephone) a good physical condition, neuropsychological impairment had not remarkably progressed. The clinical findings, and the detection of anti-Ri antibodies in serum together with the exclusion of other differential diagnosis findings are very consistent with paraneoplastic limbic encephalitis (PLE) [2]. The association of anti-Ri antibodies and carcinoid tumour is new, though not surprising, as bronchial carcinoids belong to the same histological family as small cell lung cancers (SCLC), detected in 5/28 patients with anti-Ri syndrome [5] and in 20/50 patients with PLE [2]. However, anti-Ri antibodies were not found in the latter series and have been described in only one patient with PLE before [5]. The fact that the neurological symptoms arose after an eight year latency to the diagnosis of the primary tumour may be surprising. Unfortunately we do not have a serum sample from the time before the neurological disease developed. Therefore, we cannot estimate the beginning of the anti-Ri immune response. However, as tumour recurrence was found around the time of the beginning of neurological symptoms it may be speculated, that coexpression of other factors (e. g. MHC molecules) in addition to the Ri-antigen by metastases may generate the antineuronal immune response. Such a mechanism is suggested for the anti-Hu syndrome [1]. In conclusion, this is the first reLETTER TO THE EDITORS