In a study examining the possible therapeutic effects of recombinant human tumor necrosis factor-alpha (rHuTNF-alpha) on malignant gliomas without expression of tumor necrosis factor (TNF)-receptors, RG-2 glioma cells were tested in vitro as well as in a rat experimental glioma model. A growth inhibition assay revealed no inhibiting effect in vitro up to a concentration of 20 micrograms/ml rHuTNF-alpha. Receptor-binding studies showed that RG-2 cells did not present specific receptors for rHuTNF-alpha. The pharmacokinetics of rHuTNF-alpha after intravenous injection were studied with respect to serum, tissue, and brain tumor concentrations and showed increased glioma concentrations of (mean +/- standard error of the mean) 0.47 +/- 0.18 ng TNF/mg brain compared to 0.15 +/- 0.05 ng TNF/mg brain in the normal contralateral hemisphere. No therapeutic effect on solid RG-2 gliomas could be observed after stereotactic injection of 7.3 micrograms rHuTNF/10 microliter buffer solution into the tumor in 10 animals. Immunohistochemical studies after stereotactic injection of rHuTNF-alpha showed total disappearance of the substance after 24 hours without internalization into tumor cells. Stereotactic injection of 7.3 micrograms rHuTNF 10 microliters into normal brain resulted in marked inflammatory response around the injection track, including microvascular thrombosis. These results demonstrate that rHuTNF has neither direct nor indirect cytotoxic activity on RG-2 glioma cells. Furthermore, before clinical use of rHuTNF-alpha in malignant gliomas, the authors suggest that receptor studies be done in each patient. In receptor-positive patients undergoing treatment with rHuTNF-alpha, precautions should be taken to prevent local encephalitic reactions.