In order to investigate GM2 expression in gliomas, the GM2-positive human glioma cell line (HGL) D-54 MG, which contains 0.6 nmol GM2/mg protein, representing 77% of the total monosialoganglioside fraction, was used as an immunogen for the production of anti-GM2 monoclonal antibodies. For ganglioside designations, see IUPAC-IUB (Eur. J. Biochem., 79: 11-21, 1977) and Svennerholm (J. Neurochem., 10: 613-623, 1963). Five IgM monoclonal antibodies (DMAb-1 through DMAb-5) specifically recognizing the GalNAc beta1-4(NeuAc alpha 2-3)Gal-terminal epitope common to GM2 and GalNAC-GD1a are reported. The antibodies did not react with GM1, GM3, GD2, GD3, GD1a, GD1b, and GQ1b. Purified anti-GM2 MAbs were used to define the expression of the "GM2" terminal epitope by cultured human malignant and normal cells by radioimmunoassay and membrane immunofluorescence. Among neuroectodermal tissue-derived cell lines, DMAb-3, at an optimal concentration of 5 micrograms/ml, showed high reactivity (radioimmunoassay binding ratios greater than 20) with 9 of 19 HGLs, 3 of 5 medulloblastoma, 4 of 5 neuroblastoma, and 1 of 3 melanoma lines. Moderate reactivity (binding ratio, 10-20) was exhibited by 3 HGL, 2 medulloblastoma, and 1 neuroblastoma lines and low reactivity (binding ratio, 3-10) by 5 HGL lines; no reactivity was detected with 2 HGL and 2 melanoma lines. Densitometric evaluation of monosialoganglioside extracts from human glioma and medulloblastoma cell lines in conjunction with immunostaining on thin-layer chromatograms showed that GM2 represents the major monosialoganglioside in 8 of 10 HGL and in 3 of 4 Med lines. In these lines the amount of GM2 ranged from less than 0.1 to 0.6 nmol/mg protein. These results indicate that GM2 represents a proportionally increased ganglioside of most glioma, medulloblastoma, and neuroblastoma cells in vitro.