Aldose reductase (AR) has been implicated in the etiology of diabetic cataract, as well as in other complications. However, the role of AR in these complications remains controversial because the strongest supporting evidence is drawn from the use of AR inhibitors for which specificity in vivo cannot be ascertained. To settle this issue we developed transgenic mice that overexpress AR in their lens epithelial cells and found that they become susceptible to the development of diabetic and galactose cataracts. When the sorbitol dehydrogenase-deficient mutation is also present in these transgenic mice, greater accumulation of sorbitol and further acceleration of diabetic cataract develop. These genetic studies demonstrated convincingly that accumulation of polyols from the reduction of hexose by AR leads to the formation of sugar cataracts.Diabetic complications such as neuropathy, nephropathy, retinopathy, and cataract, etc., occur in both insulin-dependent and noninsulin-dependent diabetes mellitus. Hyperglycemia has long been suspected as the cause of these manifestations, and the results of the Diabetic Control and Complications Trial (1) appear to confirm it. However, by what mechanism elevated blood glucose leads to these complications is unclear. One theory implicates the polyol pathway as a cause of diabetic cataract because of the discovery of polyols in cataractous lenses (2) and the identification of aldose reductase (AR) that reduces various sugars to their polyols (3, 4). AR reduces glucose to sorbitol, which is then converted to fructose by sorbitol dehydrogenase (SorD). Because sorbitol does not readily diffuse out of cells and its oxidation to fructose is slow, the accumulation of sorbitol under the hyperglycemic state would increase the intracellular osmotic pressure, leading to swelling and eventual rupture of the lens fiber cells (5). The involvement of AR in diabetic cataract is supported by the fact that animals such as rats and dogs that have high levels of this enzyme in their lenses are prone to develop diabetic cataract, whereas mice that have low lens AR activity are not (6). Rats and dogs also develop galactose-induced cataracts more readily than diabetes-induced cataracts (7). This fact agrees with the polyol model because galactose is a better substrate than glucose for AR in vitro, and its reduction product galactitol is not further converted to other metabolites, resulting in faster buildup of this polyol. Additional evidence for the polyol model came from the fact that several AR inhibitors could suppress cataract formation in experimentally induced diabetic animals (8-10). However, these drugs may inhibit AR by nonspecific hydrophobic interactions (11,12), and their beneficial effects may be derived from the inhibition of other enzymes. The strongest challenge to the polyol model is the fact that kinetic analyses (13, 14) and x-ray crystallographic studies (15) indicated that AR has a very low affinity for glucose and galactose, and it has not been demonstrated directly that...