BACKGROUND--Frusemide can inhibit various indirectly acting bronchoconstrictor stimuli in asthmatic patients. Both frusemide and bumetanide also modulate airway neurotransmission in some species but there are no data on the effect of loop diuretics on neurotransmission in man. An in vitro study was performed in human airways to investigate the possible neuromodulatory action of two loop diuretics, frusemide and bumetanide, and to elucidate whether a cyclooxygenase inhibitor such as indomethacin could modulate the effect of frusemide. The effect of acetazolamide, a carbonic anhydrase inhibitor, was also investigated. METHODS--Electrical field stimulation (EFS; 40 V, 0.5 ms, 0.5-32 Hz for 15 seconds) in human airways with or without epithelium was used to induce a cholinergic contraction (n = 5 in all experiments). Indomethacin was present throughout. After obtaining a control frequency-response curve, different concentrations of diuretic were added to the organ bath and another frequency-response curve was constructed. To determine whether the effect of the diuretic was prejunctional or postjunctional a cumulative concentration-response curve to exogenous acetylcholine (Ach, 0.3 mumol/l to 10 mmol/l) was constructed in the presence of a diuretic (frusemide 1 mmol/l or bumetanide 0.1 mmol/l) or its vehicle. In some experiments indomethacin was omitted from the organ bath to investigate the possible involvement of cyclooxygenase products. RESULTS--Both frusemide (10 mumol/l to 1 mmol/l) and bumetanide (1 mumol/l to 0.1 mmol/l) produced a concentration-dependent inhibition of the EFS-induced cholinergic contraction in human airways in vitro but only in epithelium denuded tissues. Frusemide (1 mmol/l) produced a maximum inhibition of 46.3% (SE 9.9%) at 0.5 Hz and bumetanide (0.1 mmol/l 39.6 (6.2)% at 0.5 Hz. Without indomethacin in the organ bath the frusemide-induced inhibition was enhanced at 4, 8, and 16 Hz, but bumetanide-induced inhibition was not enhanced at any frequency when indomethacin was omitted. Frusemide (1 mmol/l) and bumetanide (0.1 mmol/l) had no effect on the cumulative concentration-response curve to exogenous Ach (0.3 mumol/l to 10 mmol/l). Acetazolamide (100 mumol/l) had no effect on the EFS-induced cholinergic contraction in tissues with or without epithelium. CONCLUSIONS--In human airways in vitro both frusemide and bumetanide produced a concentration-dependent inhibition of the EFS-induced cholinergic contraction. This inhibition is mediated through a prejunctional mechanism. Epithelium removal was necessary to achieve this effect. The mechanism of action of frusemide and bumetanide on airway nerves remains unclear: inhibition of the Na-K-Cl cotransporter is a possibility and, for frusemide, release of endogenous cyclooxygenase products may be involved. Carbonic anhydrase inhibition, on the other hand, is unlikely to be a factor.