Immunocytochemical study of the glial fibrillary acidic protein in human neoplasms of the central nervous system.

The distribution of the glial fibrillary acidic protein (GFAP) was investigated in sections of 131 paraffin-embedded brain neoplasms obtained at surgery or at autopsy. The unlabeled antibody immunoperoxidase (peroxidase-antiperoxidase, PAP) method was used. Equally good results were obtained from 17-year-old material and from recent material derived at surgery or autopsy and fixed with Bouin fluid or phosphate-buffered formalin. The perikaryons and processes of reactive astrocytes showed the most intense stain for GFAP. Positive reaction to antibody against GFAP of varying intensity was demonstrated in astrocytomas of various grades of malignancy (32 of 32), glioblastoma multiforme (10 of 10), subependymal giant cell astrocytoma (1 of 1), ependymoma (2 of 10), subependymoma (4 of 4), and astrocytes in mixed neoplasms (8 of 8). In two neoplasms diagnosed as malignant astrocytomas and in four neoplasms diagnosed as glioblastoma multiforme, GFAP stain was limited to a few neoplastic cells. Usually the stain was more intense over processes than in perikaryons, with the exception of gemistocytic astrocytomas and the giant cells in glioblastoma multiforme, which showed an equally intense stain over perikaryons and processes. The periphery of Rosenthal fibers was intensely positive for GFAP. In astrocytic neoplasms the number of GFAP-positive cells and the intensity of the stain were inversely proportional to the degree of malignancy. In the following neoplasms the reaction for GFAP was negative: oligodendroglioma (3), oligodendroblastoma (1), medulloblastoma (3), medulloepithelioma (1), neuroblastoma (1), pineocytoma (1), typical teratoma of the pineal (1), fibrosarcoma (1), pituitary adenoma (2), craniopharyngioma (1), chordoma (1), chemodectoma of globus jugulare (1), metastatic carcinoma (17), and lymphoma (8). In one of 18 meningiomas, endogenous peroxidase activity was seen in mast cells. All meningiomas studied were negative for GFAP. In one of six neurinomas a positive reaction for GFAP was detected over processes. The authors concluded that the immunostain for GFAP is useful in the diagnoses of astrocytic neoplasms and of mixed gliomas.