Differential expression of glial- and neuronal-associated antigens in human tumors of the central and peripheral nervous system

Abstract
The immunoreactivity of a panel of poly-and monoclonal antibodies raised against different glial and neuronal antigens was investigated in paraffin-embedded specimens of 116 human tumors of the central and peripheral nervous system. We used antibodies against the HNK-1 epitope, which is shared between natural killer cells and the nervous system, glial fibrillary acidic protein (GFAP), vimentin, neurofilaments, S-100 protein, neuron-specific enolase (NSE) and myelin basic protein (MEP). HNK-1 immunoreactivity was detectable in nearly all neuroectodermal tumors. Especially in those derived from the neuroepithelium, which include the various types of gliomas, we observed a strong staining with this antibody. The only exceptions were the choroid plexus papillomas and individual ependymomas. In tumors derived from the neural crest HNK-1 reactivity was more variable and less intense. In other tumors of the nervous system HNK-1 was not detectable, except for two out of four malignant lymphomas. In addition to its reactivity with human lymphocytes HNK-1, therefore, seems to be a useful ‘marker’ for neurogenic tumors in general. GFAP expression was prominent in all astrocytomas and the astrocytic cells within mixed gliomas and gangliogliomas. Immunoreactivity was more variable in glioblastomas and ependymomas, while only isolated GFAP-positive cells were present in oligodendrogliomas, medulloblastomas, one plexus papilloma, and some neurinomas. Vimentin immunoreactivity was found in tumor cells of nearly all tumors of the central nervous system with the exception of oligodendrogliomas, most plexus papillomas, neuronal tumors and most medulloblastomas. Meningeomas, neurinomas and malignant melanomas were always strongly vimentin positive. Besides the tumor cells the vimentin antibody usually stained vascular elements within each tumor. Sarcomatous components in glioblastomas and desmoplastic areas in medulloblastomas were also labeled. Neurofilament expression was restricted to neuronal tumor cells in two gangliogliomas and to individual tumor cells in one medulloblastoma. The NSE antiserum showed more widespread and sometimes diffuse reactivity and, therefore, seems to be less valuable as an indicator for neuronal differentiation than neurofilaments. S-100 expression was demonstrable in numerous tumors including most gliomas and all tumors derived from the neural crest. MBP antibodies never showed reactivity with oligodendroglioma or neurinoma tumor cells. This antibody labeled only myelin sheaths and their remnants within these tumors. Based upon our results we can conclude that, despite the fact that most of the investigated antigens showed a widespread distribution among different tumors, each of them and especially their differential expression might be useful in the classification and differential diagnosis of human tumors of the nervous system.