Selective inhibition of phosphodiesterase (PDE) isozymes has been shown to inhibit inflammatory cell function and relax airway smooth muscle and, thus, may be useful in the therapy of asthma. In guinea pigs sensitized to ovalbumin (OA), the effects of three PDE inhibitors were compared: siguazodan (PDE III selective, IC50 = 0.7 microM), rolipram (PDE IV selective, IC50 = 0.8 microM) and zardaverine (dual PDE III/IV, IC50S = 2.5 microM and 1.1 microM, respectively) against histamine-, leukotriene (LT) D4- and OA-induced bronchospasm in vitro and in vivo. Rolipram or zardaverine (0.1-10 microM), but not siguazodan, inhibited OA-induced contraction of the isolated trachea in a concentration-dependent manner. Rolipram or siguazodan alone (10 microM) were ineffective against histamine- or LTD4-induced contractions. Zardaverine alone (10 microM) or the combination of rolipram and siguazodan (10 microM each) markedly antagonized the contractions elicited by both spasmogens. In anesthesized, ventilated guinea pigs, the i.v. ID50S against OA-induced bronchospasm were: rolipram = 0.2 mg/kg, siguazodan > 10 mg/kg and zardaverine = 2.4 mg/kg. When administered at doses up to 7.5 mg/kg, i.v., rolipram or siguazodan were markedly less effective (i.e., < or = 50% inhibition) than zardaverine (ID50S = 2.4 and 1.7 mg/kg, respectively) at blocking exogenous histamine- or LTD4-induced bronchospasm. However, when administered in combination with siguazodan (5.4 mg/kg, i.v.), rolipram (0.4-5.4 mg/kg) abolished histamine- and LTD4-induced bronchoconstriction. In conscious guinea pigs, zardaverine (5 mg/kg, intragastrically (i.g.) or the combination of rolipram and siguazodan (5 mg/kg each) were substantially more effective than rolipram or siguazodan alone at inhibiting aerosol histamine- or LTD4-induced bronchospasm. In the same animals, rolipram or zardaverine (5 mg/kg, i.g.) but not siguazodan (5 mg/kg, i.g.) markedly inhibited aerosol OA-induced bronchoconstriction. The OA-induced pulmonary eosinophil infiltration in these animals was attenuated by all treatments with zardaverine producing the greatest degree of inhibition. These results indicate that 1) PDE IV inhibitors but not PDE III inhibitors are effective at blocking antigen-induced bronchospasm, 2) compounds that selectively inhibit either PDE III or PDE IV are poor inhibitors of bronchoconstriction elicited by exogenously administered spasmogens, and 3) the combined inhibition of both PDE III and PDE IV isozymes acts in an additive or synergistic manner to inhibit bronchospasm in the guinea pig.