As determined by vanadate-facilitated [3H]ouabain binding to intact samples, semistarvation and untreated streptozotocin- or partial pancreatectomy-induced diabetes reduced rat soleus muscle Na+-K+- adenosinetriphosphatase (Na+-K+-ATPase) concentration by 12-21% (P < 0.05). Conversely, insulin treatment of rats with streptozotocin-induced diabetes induced an increase of 18-26% above control (P < 0.05). Treadmill training diminished the reduction in muscle [3H]ouabain binding site concentration induced by untreated diabetes to only 2-5%. No significant variation was observed in rat cerebral cortex Na+-K+-ATPase concentration as a result of diabetes, semistarvation, or insulin treatment. In human subjects, Na+-K+- ATPase concentration in vastus lateralis muscle biopsies was 17 and 22% greater (P < 0.05), respectively, in patients with treated non-insulin- dependent diabetes mellitus (n = 24) and insulin-dependent diabetes mellitus (n = 7) than in control subjects (n = 8). A positive linear correlation between muscle Na+-K+-ATPase and plasma insulin concentrations was observed (r = 0.50, P = 0.006; n = 29). Thus, insulin seems a regulator of muscle Na+-K+-ATPase concentration, reduction of muscle Na+-K+-ATPase concentration with untreated diabetes bears similarities with undernourishment, and physical conditioning may ameliorate the muscle Na+- K+-ATPase concentration decrease induced by diabetes.