SUMMARY: Glycosaminoglycans (GAGs) and proteoglycans of the trabecular meshwork may be involved in the pathogenesis of primary open angle glaucoma and of corticosteroid-induced glaucoma. Although numerous studies of GAGs have been performed on trabecular cells in monolayer culture and in animals, few studies have examined intact human meshwork. Using a new analytic method for quantitative determination of GAGs from individual meshworks, we studied the GAGs of human trabecular meshwork in normal eyes and intact meshwork from perfusion-cultured eyes. Eyes in the culture system were treated with dexamethasone for 21 days in an attempt to create a model of steroid glaucoma. The effect of ascorbate, also known to influence GAG levels, was examined in a preliminary study. Cultured trabecular meshwork had a significant decrease in total GAG levels when compared with meshwork from normal, noncultured eyes (1.16 +/- 0.54 ng/mug vs. 3.51 +/- 0.57 ng/mug; p < 0.01). Ascorbate tended to cause a decrease (28%) in the total level of GAGs, although the difference was not significant. No significant difference in mean GAG levels was found between the dexamethasone-treated and fellow control eyes. Dexamethasone was associated with a large biologic variability in total GAG level (coefficient of variation 89%). In an attempt to identify individual steroid responders mixed in with a group of nonresponders, analysis of individual eyes revealed three of nine eyes to have total GAGs 250% than fellow control eyes, although this variability could also be due to other causes. No consistent change in group mean intraocular pressures was found in the dexamethasone-treated eyes. Intraocular pressures were positively correlated with total GAG levels (r = 0.59; p < 0.01).