It has been proposed that increased supply of gluconeogenic precursors may be largely responsible for the increased gluconeogenesis which contributes to fasting hyperglycemia in non-insulin-dependent diabetes mellitus (NIDDM). Therefore, to test the hypothesis that an increase in gluconeogenic substrate supply per se could increase hepatic glucose output sufficiently to cause fasting hyperglycemia, we infused normal volunteers with sodium lactate at a rate approximately double the rate of appearance observed in NIDDM while clamping plasma insulin, glucagon, and growth hormone at basal levels. In control experiments, sodium bicarbonate was infused instead of sodium lactate at equimolar rates. In both experiments, [6-3H]-glucose was infused to measure glucose appearance and either [U-14C]lactate or [U-14C]alanine was infused to measure the rates of appearance and conversion of these substrates into plasma glucose. Plasma insulin, glucagon, growth hormone, C-peptide, and glycerol concentrations, and blood bicarbonate and pH in control and lactate infusion experiments were not significantly different. Infusion of lactate increased plasma lactate and alanine to 4.48 +/- 3 mM and 610 +/- 33 microM, respectively, from baseline values of 1.6 +/- 0.2 mM and 431 +/- 28 microM, both P less than 0.01; lactate and alanine rates of appearance increased to 38 +/- 1.0 and 8.0 +/- 0.3 mumol/kg per min (P less than 0.01 versus basal rates of 14.4 +/- 0.4 and 5.0 +/- 0.5 mumol/kg per min, respectively). With correction for Krebs cycle carbon exchange, lactate incorporation into plasma glucose increased nearly threefold to 10.4 mumol/kg per min and accounted for about 50% of overall glucose appearance. Alanine incorporation into plasma glucose increased more than twofold. Despite this marked increase in gluconeogenesis, neither overall hepatic glucose output nor plasma glucose increased and each was not significantly different from values observed in control experiments (10.8 +/- 0.5 vs. 10.8 +/- 0.5 mumol/kg per min and 5.4 +/- 0.4 vs. 5.3 +/- 0.3 mM, respectively). We, therefore, conclude that in normal humans there is an autoregulatory process independent of changes in plasma glucose and glucoregulatory hormone concentrations which prevents a substrate-induced increase in gluconeogenesis from increasing overall hepatic glucose output; since this process cannot be explained on the basis of inhibition of gluconeogenesis from other substrates, it probably involves diminution of glycogenolysis. A defect in this process could explain at least in part the increased hepatic glucose output found in NIDDM.