We have evaluated the effects of the novel immunosuppressant sodium fusidate (fusidin) in the non-obese diabetic (NOD) mouse and in D-galactosamine (D-Gal)-presensitized BALB/c mice challenged with the bacterial superantigen, Staphylococcus aureus enterotoxin B (SEB) or with the endotoxin, Escherichia coli lipopolysaccharide (LPS). The NOD mouse model has clinical and histoimmunological features similar to those of human insulin-dependent diabetes mellitus (IDDM). The SEB- and LPS-treated BALB/c mouse models exhibit pathogenic similarities with human septic shock conditions. In the NOD mouse, fusidin suppressed the spontaneous development of insulitis (mean inhibition 73%) and hyperglycaemia (IDDM incidence 25% versus 0%) when administered at 40 mg/kg five times weekly for 8 consecutive weeks from the fourth week of age; concurrently treated animals exhibited reduced percentages of splenic T lymphocytes. This anti-diabetogenic effect was confirmed in the accelerated model of diabetes induced in the NOD mouse with cyclophosphamide (CY) (IDDM incidence 55% versus 21-6% using dosages of fusidin from 40 to 80 mg/kg five times weekly); protection from IDDM development was achieved even when the drug (80 mg/kg/day) was first administered 7 days after CY challenge. In contrast, fusidin did not reverse hyperglycaemia when administered to CY-treated animals within 3 days of IDDM development. In the two models of septic shock, prophylactic treatment with fusidin, 80 mg/kg given three times for 2 days prior to D-Gal/SEB or D-Gal/LPS challenge, drastically reduced the lethality compared with D-Gal/buffer-treated mice. This effect may depend on the inhibitory action of fusidin on the secretion of cytokines such as interferon-gamma and tumour necrosis factor-alpha, the serum levels of which were greatly diminished in the fusidin-treated mice (mean inhibition 50-90%). These results demonstrate that fusidin may have a role in the treatment of cell-mediated autoimmune diseases and cytokine-mediated infectious diseases in humans.