Diabetes and the myo-inositol paradox.
 To test the general applicability of the hypothesis that diabetes mellitus causes increased polyol pathway activity, decreased tissue free myo-inositol, and resultant pathological changes in tissues susceptible to the ravages of diabetes, we measured glucose, sorbitol, and myo-inositol with quantitative histochemical techniques in layers of the cornea, the aortic myointima, the cardiac left ventricle and atrioventricular node (AVN), and retina and kidney after 19 days or 2 mo (mildly diabetic non-insulin-treated [MD] and severely diabetic insulin-treated [SD] groups) in the alloxan-induced diabetes model.
 In the aqueous humor, glucose rose linearly with increased serum glucose, sorbitol was markedly increased in the MD and SD groups, and myo-inositol did not change in any diabetic group.
 There was no change in glucose or sorbitol in aortic myointima in any group, but myoinositol was decreased in 19-day diabetic rabbits by 26%, unchanged in MD rabbits but paradoxically increased by 60% in SD rabbits.
 Glucose, sorbitol, and myo-inositol increased in all three corneal layers in SD rabbits but only in epithelium and stroma in 19-day and MD rabbits.
 AVN glucose and sorbitol did not change in 19-day diabetic, MD, or SD diabetic rabbits.
 AVN myo-inositol was three times higher than ventricular myo-inositol and did not appear to change in SD rabbits.
 Retinal pigmented epithelium myo-inositol was decreased 30% in SD rabbits.
 Glomerular myo-inositol was also decreased, but not significantly, in SD rabbits.
 We conclude that the paradoxical increase in corneal and aortal myo-inositol raises fundamental questions about the general applicability of the myo-inositol-depletion hypothesis.
