The purpose of the study was to develop a mouse model of non-insulin-dependent diabetes mellitus (NIDDM) that closely simulates the metabolic abnormalities of the human disease and is also cost-effective compared with the genetic models currently available. For this purpose, insulin resistance was induced in male C57BL/6J or Institute of Cancer Research (ICR) mice by feeding diets enriched in either fructose or fat, and hyperglycemia was induced by injecting these mice with a dose of streptozotocin (STZ) that does not cause diabetes in chow-fed mice. In the case of C57BL/6J mice, insulin levels initially increased in response to the fructose- and fat-enriched diets and then decreased to levels comparable to or still higher than those in chow-fed mice following STZ injection. Associated with the decrease in insulin levels following STZ, fat-fed and fructose-fed C57BL/6J mice became significantly hyperglycemic, reaching values of 388 +/- 38 and 366 +/- 58 mg/dL, respectively. In contrast, neither plasma glucose nor insulin concentrations changed in chow-fed mice injected with an identical amount of STZ. Essentially identical findings were seen before and after STZ injection in fat-fed compared with chow-fed ICR mice. Although a direct comparison was not made, sensitivity to the diabetogenic effects of STZ appeared to be greater in fat-fed ICR compared with fat-fed C57BL/6J mice. Finally, plasma glucose decreased when mice with these experimental models of NIDDM were treated with either metformin or tolbutamide. Given these results, it seems reasonable to suggest that the combination of dietary-induced insulin resistance and relatively low-dose STZ results in mouse models that should be of use in studying the pathophysiology of NIDDM or in evaluating therapeutic compounds for the treatment of NIDDM.