Previous studies suggested that circulating tumor necrosis factor alpha (TNF-alpha) may have a pathophysiologic role in experimental neonatal sepsis induced by group B streptococci (GBS). This study was undertaken to investigate the ability of the type III and group-specific polysaccharides of GBS to induce TNF-alpha production and TNF-alpha-dependent lethality in neonatal rats. The cytokine was detected in plasma samples by the L929 cytotoxicity assay. Intracardiac injections of either polysaccharide induced dose-dependent, transient elevations in plasma TNF-alpha levels that returned to baseline values after 5 h. The group-specific antigen induced significantly higher mean peak TNF-alpha levels than the type III antigen (125 +/- 47 versus 44 +/- 15 U/ml with 70 mg/kg of body weight). Glycogen (70 mg/kg), used as a negative control, did not induce TNF-alpha. The lipopolysaccharide-neutralizing agent polymyxin B did not decrease TNF-alpha levels induced by either polysaccharide, ruling out contamination with endotoxin as a possible cause of TNF-alpha induction. Fifty percent lethal doses of the type III and group-specific antigens given as intracardiac injections were 105 and 16 mg/kg, respectively. Salmonella endotoxin, used as a positive control, had a 50% lethal dose of 0.1 mg/kg. The lethal activities of GBS polysaccharides, as well as endotoxin, were completely prevented by pretreatment of neonatal rats with the respective specific antibodies or anti-murine TNF-alpha serum. To assess the relative importance of the type-specific substance in TNF-alpha induction by whole bacteria, two unrelated GBS transposon mutants devoid of only the type-specific capsular polysaccharide (COH1-13 and COH31-15) were employed. Each of the heat-killed unencapsulated mutants was able to produce plasma TNF-alpha level elevations or TNF-alpha-dependent lethality but was significantly less efficient in these activities than the corresponding encapsulated wild-type strain. These data suggest that the presence of type-specific material on GBS is not necessary for the stimulation of TNF-alpha production. Type III capsular polysaccharide, however, can significantly increase the ability of GBS to induce TNF-alpha. Further studies will be needed to assess the importance of TNF-alpha induction by the group- and type-specific antigens in the pathophysiology of GBS disease.