The mechanisms of testosterone(T)-induced, endothelium-dependent vasodilation are not fully understood. Studies showed that endothelial NO partly mediated this effect (1, 2). The production of endothelial vasodilators, e.g. NO and prostacyclin, are modulated by endothelial ion channels which influence the resting membrane potential, Ca2+ electrochemical driving force, the magnitude of Ca2+ influx and ultimately [Ca2+]i (3, 4). This study investigated the effect of T on ion channels of human coronary endothelial cells (HCAECs) and its molecular mechanism using whole-cell patch clamp technique. Values are mean ± SEM (no.) of %increase from control currents. T at 100-1000nM induced a rapid (2-5 minutes), dose-dependent increase in total current [100nM T: 31.9±8.6% (18), 200nM T: 38.8±10.9% (8), 300nM T: 59.7±15.8% (18), and 1μM T: 58.7±7.1% (37)]. The EC50 of T, obtained by curve-fitting to the Hill equation was 83.7±1.9nM. 1μMT, after pretreatment with DIDS and La3+ (to block Cl- and transient receptor potential channels, respectively) increased the residual currents (primarily carried by K+) by 45.2±9.0% (n=13, p<0.05, t-test). In contrast, 1μMT did not change currents after pretreatment with K+ channel blockers TEA + BaCl2 and DIDS or La3+ (15.8±8.7% and -0.1±14.1%, n=7 and 7, respectively, n.s., t-test). Non-permeant form of T, bovine serum albumin-conjugated T (T-BSA, 1μM), increased HCAEC currents by 93.1±10.2% (n=17, p<0.05; t-test). Pretreating with an androgen receptor antagonist flutamide reduced T effect, from 84.8±22.9% (n=7) without to 1.7±4.3% (n=8) with flutamide pretreatment, respectively (p<0.05; t-test). The effect of T-BSA was also inhibited by flutamide. However, pretreating with estrogen receptor blocker fulvestrant did not prevent T effects; 101.5±22.0% (n=3) in T vs. 75.2±11.7% (n=8) in T + fulvestrant pretreatment (n.s.; t-test). Dihydrotestosterone (DHT), a nonaromatizable androgen, also stimulated HCAEC currents, which was attenuated by flutamide pretreatment; 195.7±45.1% (n=7) without vs. 2.6±19.5% (n=4) with flutamide pretreatment (p<0.05; t-test). Incubating HCAECs with pertussis toxin (PTX) partly inhibited the HCAEC current increase by T; 63.1±16.2% (n=6) in T vs. 15.2±7.5% (n=6) in PTX pretreatment + T (p<0.05, t-test), while pre-incubation with a PLC inhibitor U-73122 did not significantly prevent the T effect. These data suggested that androgens may non-genomicly increase HCAEC K+ currents via surface androgen receptors. This effect may be partly mediated by Go/i-protein but not PLC.