Glutamine synthetase (GS, EC 6.3.1.2.) has long been considered as a protein specific for astrocytes in the brain, but recently GS immunoreactivity has been reported in oligodendrocytes both in mixed primary glial cell cultures and in vivo. We have investigated its expression and regulation in "pure" oligodendrocyte cultures. "Pure" oligodendrocyte secondary cultures were derived from newborn rat brain primary cultures enriched in oligodendrocytes as described by Besnard et al. (1987) and were grown in chemically defined medium. These cultures contain more than 90% galactocerebroside-positive oligodendrocytes and produce "myelin" membranes (Fressinaud et al., 1990) after 6-10 days in subcultures (30-35 days, total time in culture). The presence of GS in oligodendrocytes from both primary glial cell cultures and "pure" oligodendrocyte cultures was confirmed by double immunostaining with a rabbit antisheep GS and guinea pig antirat brain myelin 2', 3'-cyclic nucleotide 3'-phosphodiesterase. In "pure" oligodendrocyte cultures, about half of cells were labeled with anti-GS antibody. Furthermore, on the immunoblot performed with a rabbit antisheep GS, the GS protein in "pure" oligodendrocyte secondary cultures was visualized as a single band with an apparent molecular mass of about 43 kDa. In contrast, two protein bands for GS were observed in cultured astrocytes. On the immunoblot performed with a rabbit antichick GS, two immunopositive protein bands were observed: a major one migrating as the purified adult chick brain GS and a minor one with a lower molecular mass. Two similar immunoreactive bands were also observed in pure rat astrocyte cultures. Compared to pure rat astrocyte cultures, "pure" oligodendrocyte cultures of the same age displayed an unexpectedly high GS specific activity that could not be explained by astrocytic contamination of the cultures (less than 5%). As for cultured astrocytes, treatment of oligodendrocyte cultures with dibutyryl-adenosine 3':5'-cyclic monophosphate, triiodothyronine, or hydrocortisone increased significantly GS specific activity. Interestingly, epidermal growth factor, basic fibroblast growth factor, and platelet-derived growth factor that increase the GS activity in astrocytes do not affect this activity in oligodendrocytes. Thus we confirm the finding of Warringa et al. (1988) that GS is also expressed in oligodendrocytes. We show that its activity is regulated similarly in astrocytes and oligodendrocytes by hormones, but that it is regulated differently by growth factors in these two cell types.