Experimental and therapeutic use of islet cell preparations could benefit from assays that measure variations in the mass of living beta-cells. Because processes of cell death can be followed by depletion and/or discharge of cell-specific substances, we examined whether in vitro conditions of beta-cell death resulted in changes in tissue and medium content of insulin and of gamma-aminobutyric acid (GABA), two beta-cell-specific compounds with different cellular localization and turnover. Exposure of rat purified beta-cells to streptozotocin (5 mM, 120 min) or to the nitric oxide donor GEA-3162 (GEA; 50 microM, 120 min) caused 80% necrosis within 24 h; at the end of this period, cellular insulin content was not significantly decreased, but cellular GABA content was reduced by 70%; when cultured at basal glucose (6 mM), the toxin-exposed cells did not discharge less insulin but released 80% less GABA in the period 8-24 h. As in rat beta-cell purification, GABA comigrated with insulin during human islet cell isolation. Twenty-four hours after GEA (500 microM, 120 min), human islet cell preparations exhibited 90% dead cells and a 45 and 90% reduction, respectively, in tissue insulin and GABA content; in the period 9-24 h, insulin discharge in the medium was not reduced, but GABA release was decreased by 90%. When rat beta-cells were cultured for 24 h with nontoxic interleukin (IL)-1beta concentrations that suppressed glucose-induced insulin release, cellular GABA content was not decreased and GABA release increased by 90% in the period 8-24 h. These data indicate that a reduction in cellular and medium GABA levels is more sensitive than insulin as a marker for the presence of dead beta-cells in isolated preparations. Pancreatic GABA content also rapidly decreased after streptozotocin injection and remained unaffected by 12 h of hyperglycemia. At further variance with insulin, GABA release from living beta-cells depends little on its cellular content but increases with IL-1beta-induced alterations in beta-cell phenotype.