In the present study, we examined whether substance P (SP) and SP methyl ester (SPME), a selective NK(1) agonist, cause biphasic responses consisting of endothelium-dependent relaxation (EDR) and contraction (EDC) in precontracted rabbit intrapulmonary arteries. In arteries contracted with PGF(2alpha) (2x10(-6) M), SP as well as SPME caused only EDR at low concentration (10(-9) M) and EDR followed by EDC at higher concentrations, indicating the involvement of NK(1) receptors. The SP (10(-8) M)-induced EDR was abolished in arteries moderately contracted by PGF(2alpha) (5x10(-7) M) and the EDC in arteries maximally contracted by PGF(2alpha) (10(-5) M), indicating that EDR and EDC are inversely dependent on preexisting tone. Indomethacin (10(-8) - 10(-6) M), a cyclo-oxygenase inhibitor, and ozagrel (10(-8) - 10(-6) M), a TXA(2) synthetase inhibitor attenuated the EDC in the SPME (10(-7) M)-induced biphasic response and markedly potentiated the EDR. AA-861 (10(-8) - 10(-6) M), a 5-lipoxygenase inhibitor, did not affect the EDR or EDC. L-N(G)-nitro-arginine methyl ester (10(-5) - 10(-4) M), a nitric oxide synthase inhibitor, attenuated the EDR and slightly potentiated the EDC. CP-99994 (10(-10) - 10(-8) M), an NK(1) antagonist, attenuated the EDC and potentiated the EDR in the SPME (10(-7) M)-induced biphasic response, while the NK(2) antagonist SR-48968 (10(-9) - 10(-7) M) had no effect. CP-99994 attenuated the SPME (10(-7) M)-induced EDC under EDR-blockade to a greater extent than the EDR under EDC-blockade, indicating that CP-99994 enhanced the EDR component by preferential inhibition of the EDC component. In conclusion, NK(1) agonists caused a biphasic endothelium-dependent response (EDR and EDC) in submaximally precontracted intrapulmonary arteries. The EDC and EDR mediated by NK(1) receptors may play physiological and/or pathophysiological roles in modulation of vascular tone.