The neu/c-erbB-2 oncogene encodes a 185 kd transmembrane protein (p185). Here we have used the monoclonal antibody (mAb) 3B5 to determine the expression of p185 in a series of fixed biopsy specimens of 180 human brain tumors, including the most frequent entities and, in addition, 18 recurrent gliomas with malignant progression. In summary, 3B5 immunoreaction was most prominent in astrocytomas of different grades of malignancies and in meningiomas. In World Health Organization (WHO) grade II astrocytomas mab 3B5-immunoreaction was related to the cytomorphological phenotype. Fibrillary astrocytomas showed no or only a weak immunoreaction (four of five, 80%) in contrast with protoplasmic or gemistocytic astrocytomas, where a strong reaction was observed in most cases (six of nine, 66.6%, and four of five, 80%, respectively). In WHO grade II to WHO grade IV astrocytomas a trend towards higher scores with increasing grade was found. In a limited number of cases (18 gliomas and two meningiomas) of the tumor series tested other mAbs against neu/c-erbB-2 epitopes, especially the mabs 9G6 and CB11, gave qualitatively comparable results. In WHO grade I pilocytic astrocytomas a wide range of 3B5 immunoreactivity has been observed. The results of in situ hybridization using a 32P-labeled neu/erbB-2 RNA probe performed on four WHO grade I and II astrocytomas, seven WHO grade IV glioblastomas, one WHO grade II oligoastrocytoma, one WHO grade III anaplastic astrocytoma, and three WHO grade I meningiomas were consistent with these immunomorphological data, and Northern blot analysis also indicated an overexpression of neu/c-erbB-2 mRNA in gliomas of different grades of malignancy and in meningiomas. These elevated neu-erbB-2 transcript levels occurred in the absence of gene amplification. In a second series of recurrent gliomas with malignant progression (n = 18) the higher 3B5-immunoreaction scores were apparent in the more malignant recurrent gliomas. In this series the overexpression of neu/c-erbB-2 parallels glioma progression. In our cases it was not, however, correlated with the postoperative relapse-free interval or with the overall length of survival.