Antibody repertoire in paraneoplastic cerebellar degeneration and small cell lung cancer

Abstract

The goal of this study is to determine whether patients with paraneoplastic cerebellar degeneration (PCD) and small-cell lung cancer (SCLC) have a specific repertoire of antibodies, if SOX1 antibodies (SOX1-ab) can predict the presence of SCLC, and if antibodies to cell surface antigens occur in this syndrome. Antibody analysis was done using immunohistochemistry on rat brain, immunoblot with recombinant antigens, screening of cDNA expression libraries, and immunolabeling of live neurons in 39 patients with PCD and SCLC. VGCC-ab were measured by RIA, and SOX1-ab, Hu-ab, and ZIC4-ab by immunoblot. Lambert-Eaton myastenic syndrome (LEMS) was present in 10 of 23 patients with electrophysiological studies. At least one antibody was detected in 72% of patients. The individual frequencies were: 49% SOX1-ab, 44% VGCC-ab, 31% Hu-ab, and 13% ZIC4-ab. SOX1-ab occurred in 76% of patients with VGCC-ab and 27% of those without VGCC-ab (p = 0.0036). SOX1-ab were not found in 39 patients with sporadic late-onset cerebellar ataxia, 23 with cerebellar ataxia and glutamic acid decarboxylase antibodies, and 73 with PCD and cancer types other than SCLC (31 without onconeural antibodies, 25 with Yo-ab , 17 with Tr-ab). Five patients (13%) had antibodies against unknown neuronal cell surface antigens but none of them improved with immunotherapy. One serum immunoreacted against the axon initial segment of neurons and another serum against ELKS1, a protein highly expressed in the cerebellum that interacts with the beta4-subunit of the VGCC. In conclusion, 72% of patients with PCD and SCLC had one or more antibodies that indicate the presence of this tumor. In these patients, VGCC-ab and SOX1-ab occur tightly associated. SOX1-ab are predictors of SCLC in ataxia patients with a specificity of 100% and sensitivity of 49%. Unlike limbic encephalitis with SCLC, antibodies to cell surface antigens other than VGCC-ab, are infrequent and do not predict response to treatment.