We have shown previously that tri-iodothyronine (T3)-induced sex hormone-binding globulin (SHBG) secretion by the human hepatoblastoma cell line, HepG2, can be modulated by retinoids. We have now used this model to study a range of other compounds that are known to influence T3 responsiveness in various cell systems. HepG2 cells were incubated for 4 days in serum-free medium containing T3, together with insulin, dexamethasone, phorbol myristate (PMA), sodium butyrate or estradiol. T3 (10 nmol/l) alone induced a concentration of SHBG secreted by HepG2 cells that was 187 +/- 20% (mean +/- S.D., n = 9) of control. Insulin (100 nmol/l) reduced basal SHBG secretion from 24.7 +/- 5.2 nmol/l to 16.1 +/- 1.7 nmol/l (P < 0.01). This effect was dose responsive, half-maximal at 3.4 +/- 3.0 nmol/l (approximately 600 mU/l) and maximal with 100 nmol/l insulin. Co-incubating 0-10 nmol/l T3 with 100 nmol/l insulin resulted in a downward shift in the dose-response curve without a change in the half-maximal response to T3. Conversely, 0-100 nmol/l insulin reduced SHBG production induced by 10 nmol/l T3. In contrast; while dexamethasone alone was without effect on SHBG secretion, 100 nmol/l dexamethasone induced a shift to the left in half-maximal T3 stimulation from 0.37 nmol/l to 0.10 nmol/l. The effect of PMA on SHBG secretion was reminiscent of the previously observed retinoid effect. PMA 100 nmol/l abolished maximal T3 stimulation. This effect was dose responsive, with a threshold at 1 nmol/l PMA. Sodium butyrate, up to 1 mmol/l was without effect; with greater concentrations, SHBG secretion was reduced. T3 responsiveness was virtually abolished by 3 mmol/l sodium butyrate; higher concentrations were cytotoxic and secretion was reduced to less than 20% of basal. Lack of an effect of estradiol on SHBG secretion by HepG2 cells was confirmed. These studies suggest that T3-induced SHBG secretion by HepG2 cells is independently influenced by insulin, potentiated by dexamethasone, and modulated by PMA. Detailed molecular analysis of this model will increase our understanding of the mechanism of action of T3, specifically in human liver cells.