Severe pulmonary hypertension is a fatal disease with short life expectancy 12. Continuous intravenous administration of prostacyclin was documented to improve exercise capacity and survival in patients with idiopathic pulmonary arterial hypertension (IPAH, formerly primary pulmonary hypertension, PPH) 13. Possible disadvantages of this approach are catheter related septic events and systemic side effects including serious systemic hypotension. In patients with pulmonary hypertension associated with pulmonary fibrosis, systemic administration of vasodilators results in ventilation perfusion mismatch and impairment of arterial oxygenation. Inhalation of aerosolized iloprost, a long-acting prostacyclin analogue, has been shown to cause selective pulmonary vasodilatation in both primary and secondary pulmonary hypertension 456. Long term use of nebulized iloprost was described to improve exercise capacity, event-free survival and hemodynamics in severe IPAH, and this finding was supported by a randomized, controlled phase III study in patients in NYHA class III and IV 7.

The use of aerosolized iloprost for acute pulmonary vasodilatation and putative long-term anti-remodeling effects in severe chronic pulmonary hypertension does, however, demand 6 – 12 inhalations per day, as the vasodilatory effect levels off within ~60 min post nebulization. Against this background, recent studies addressed the impact of selective phosphodiesterase (PDE) inhibitors on prostacyclin-induced acute pulmonary vasodilatation, reporting a marked amplification and prolongation of the vasodilatory response to inhaled PGI₂ 8. PDEs are enzymes that inactivate cyclic AMP and cyclic GMP, the second messengers of prostacyclin and NO 910. The characterization of the various PDEs currently known has largely profited from the employment of selective PDE inhibitors. Concerning the lung vasculature, the presence of the PDE isoenzymes 1, 3, 4 and 5 in the cytosolic and particulate phases (homogenized human pulmonary artery tissue) has been demonstrated 11.

Phosphodiesterase 1 is Ca²⁺/calmodulin dependant and hydrolyzes both cGMP and cAMP. PDE3 does possess high affinity for both cAMP and cGMP, with V_max for cAMP usually greater than that for cGMP 912. PDE4 enzymes are characterized by their high affinity to cAMP, with cGMP representing a very poor substrate. In contrast, PDE 5 is cGMP-specific and was found to be highly expressed in lung tissue 1314.

Recent clinical data suggest that the PDE 1/5/6 inhibitor sildenafil (IC₅₀ values 280 nM, 3.5 nM and 37 nM, respectively 15), which has been approved for the treatment of erectile dysfunction, is an effective pulmonary vasodilator in patients with pulmonary arterial hypertension 161718192021. Based on a very recent positive phase III study, sildenafil has been approved for the treatment of pulmonary hypertension in US. Interestingly, it has been shown that sildenafil synergizes with inhaled iloprost in patients with pulmonary hypertension 1622. Hitherto no attempt was undertaken to clarify, which of the PDEs addressed by sildenafil is the most relevant for the effect of this agent in the pulmonary circulation, and whether combinations with further selective PDE inhibitors might even enhance the sildenafil effect. To address this issue, systemic application of per se ineffective doses of specific PDE inhibitors in companion with inhalation of iloprost was undertaken in an experimental model of pulmonary hypertension in the present study.

The inhalation of nebulized prostanoids for treatment of pulmonary hypertension is an approach targeting selective vasodilatation in well ventilated and aerosol accessible lung regions. This strategy has been developed to circumvent side effects of the conventional intravenous therapy with prostacyclin, e.g. systemic hypotension and ventilation-perfusion mismatch. As anticipated from previous studies in patients with severe pulmonary hypertension 45, aerosolization of iloprost was, indeed, effective in causing lung vasorelaxation without any decrease in systemic arterial pressure and any deterioration of gas exchange in the present rabbit model. After completion of the aerosolization maneuver, the iloprost effect levelled off within 20 – 30 min, which is somewhat more rapid than in the human system (45–90 min). This difference is most likely due to some species variance in the kinetics of the iloprost catabolic pathway: being chemically stable – in contrast to prostacyclin – iloprost is converted to dinor-and tetranor-iloprost metabolites via beta-oxidation 25. The liver is known to be a major site of this catabolic pathway, however, recent studies in isolated perfused rabbit lungs demonstrated that the conversion of iloprost to these beta-oxidation products also takes place in the lung tissue itself 26.

Phosphodiesterases represent the major route for cAMP and cGMP degradation in cells, thereby limiting the downstream effects of adenylate and guanylate cyclase activating agents such as prostacyclin or NO. Until now, 11 types of PDEs have been characterized, which differ in substrate specificity and regulatory properties 92728. Within the lung, PDE 3 and 4 represent the major cAMP hydrolyzing pathways 10, and monoselective inhibitors of PDE 3 and PDE 4 have been shown to possess pulmonary vasodilatory potency 81129.

Motapizone is a highly selective PDE3 inhibitor, with an IC₅₀ value of 30 nM 30. Thus, the finding that motapizone infusion causes dose-dependent pulmonary vasodilatation in intact rabbits with elevated pulmonary artery pressure is well in line with previous observations in this field. Notably, the motapizone effect was not pulmonary selective: in the dose range from 10 to 1000 μg/kg × min, both the pulmonary artery and the systemic artery pressure declined in a parallel fashion.

PDE1 gene products are expressed in cardiac tissues from several species 3132. A role of increased PDE1C expression in the cardio-protective effect of the stable prostacyclin derivative, 7-oxo-prostacyclin, indicates that PDE1C variants may be involved in tissue responses to cardiovascular stress 33. In vascular smooth muscle cells derived from different species, several reports demonstrated PDE1 expression. PDE1C was shown to be present in proliferating smooth muscle cells 3134 and an increased expression of PDE1A1 in rat aorta was shown to contribute to the development of nitroglycerin tolerance 35.

Concerning the lung vasculature, only very limited data on PDE1 expression is available 11. Our group recently observed that PDE1C is strongly upregulated in the pulmonary artery media of human lungs with severe pulmonary hypertension (R. Schermuly et al., non-published results). It is in line with this notion that the selective PDE1 inhibitor 8MM-IBMX induced dose-dependent pulmonary vasodilation in the presently investigated acute pulmonary hypertension model, however, without being specific for the pulmonary circulation, as evident from the parallel decline of systemic arterial pressure.

The cGMP-specific phosphodiesterase PDE 5 is abundantly distributed in the lung tissue 131436. In a hypoxia-induced model of pulmonary hypertension in the rat, Cohen et al. demonstrated that the PDE 5 inhibitor E4021 selectively vasodilates the pulmonary circulation when being applied intravenously 36, and this observation was confirmed in a model of newborn lambs with persistent pulmonary hypertension 37. Another specific PDE 5 inhibitor, E4010, has been shown to be a selective pulmonary vasodilatator in a hypoxic rat model of pulmonary hypertension 38. Accordingly, the PDE 1/5/6 inhibitor sildenafil, which is approved for treatment of erectile dysfunction, was also recently shown to cause preferential pulmonary vasodilatation even when being systemically administered 39. These data are well in line with the current finding in pulmonary hypertensive rabbits that intravenously infused sildenafil causes a doses-dependent decrease in pulmonary artery pressure, virtually without any decline in systemic arterial pressure.

The rationale to combine cAMP-elevating agents, like prostacyclin or the stable prostacyclin analogue iloprost, with PDE3 inhibitors is obvious, and studies from our group already showed a marked amplification and prolongation of the pulmonary vasodilatory response to aerosolized prostacyclin in the presence of type 3 PDE inhibitors 829. The present investigation demonstrates that such synergistic effect also hold true for the PDE3 inhibitor motapizone and the longer acting agent iloprost: in the presence of minute doses of intravenously applied motapizone, the area under the curve of P_PA decrease in response to the standard inhaled iloprost dose was nearly threefold increased, again without any decline in systemic arterial pressure or any deterioration of gas exchange being detectable.

Interestingly, similar potency to increase the response to inhaled iloprost was also noted for subthreshold doses of the PDE 1 inhibitor 8MM-IBMX. This might be anticipated to some extent, as PDE1 causes degradation of both cAMP and cGMP (Fig. 6), thereby directly effecting downstream signalling of iloprost, and indirectly modifying this pathway via cGMP sensitive PDE inhibitors. The fact, however, that the area under the curve of PPA reduction was similarly augmented as in the presence of motapizone (~threefold) suggests that PDE1 inhibitors are worth to be taken into consideration for further strategies to enhance beneficial prostanoid effects in the pulmonary circulation.

Furthermore, the current study demonstrates that the PDE 1/5/6 inhibitor sildenafil also amplifies the pulmonary vasodilatory response to inhaled iloprost, and that in this respect subthreshold systemic doses of sildenafil are again virtually as effective as subthreshold doses of the PDE3 inhibitor motapizone. This amplification of the iloprost-induced P_PA decrease again occurred without any decline in systemic arterial pressure and any gas exchange disturbances. The mechanisms underlying this sildenafil effect deserve further elucidation. Given the pharmacological profile of this agent 4041, it is unlikely that sildenafil caused relevant direct inhibition of lung PDE 3 and 4, thereby promoting prolongation of the half life of cAMP. Its effect may, however, well be explained by the known cross talk between the cAMP and the cGMP pathways: PDE 3 is inhibited by intracellular cGMP with an IC₅₀ of 0.1–1 μM 42. A significant inhibition of PDE 5 by sildenafil may thus result in cGMP accumulation, given the fact that there is some permanent baseline stimulation of this pathway via endogenous NO, and next to causing per se some vasodilatory effect, the sildenafil-induced cGMP may particularly be effective via PDE 3 inhibition and thereby enhanced sensitivity to inhaled iloprost. In addition to such mode of action, supported by previous studies in the cooperativity between the cGMP and the cAMP axis in the lung vasculature 843, further (non-cGMP related) effects of sildenafil in the pulmonary circulation may involve inhibition of PDE1 (Fig. 5). This view is supported by the above discussed efficacy of selective PDE1 inhibition by 8-MM-IBMX. The IC₅₀ of sildenafil against PDE1 is about 280 nM 15, and although plasma levels of sildenafil are not addressed in this study, these levels could be achieved after sildenafil application.

Thus, both indirect inhibition of PDE3 by increased cGMP and direct inhibition of PDE1 may explain the effects of sildenafil on the iloprost-induced vasodilation. Our studies in co-application of sildenafil with motapizone on the one hand and 8-MM-IBMX on the other hand do, however, favour the sildenafil-PDE1 axis: whereas the combination of sildenafil with the PDE1 inhibitor 8MM-IBMX did not further amplify the iloprost-induced vasodilation over the effect of each agent alone, the combination of sildenafil plus motapizone effected a further strong amplification of the prostanoid induced vasodilation. Besides being of interest as to the mode of action of sildenafil, this finding suggests that an optimum strategy to combine PDE inhibitors may result in even further augmentation of pulmonary vascular prostanoid efficacy as compared to the choice of one selective PDE inhibitor as partner for the prostanoid. In conclusion, intravenous administration of the PDE1 inhibitor 8MM-IBMX, the PDE 3 inhibitor motapizone and the PDE 1/5/6 inhibitor sildenafil causes dose-dependent pulmonary vasodilation in a rabbit model of pulmonary hypertension, with sildenafil possessing selectivity for the lung vasculature. Most interestingly, when applied in subthreshold doses, all PDE inhibitors enhanced and markedly prolonged the vasodilatory response to inhaled iloprost, without any systemic pressure decline or deterioration of gas exchange being detectable. Maximum efficacy was noted upon combination of sildenafil and motapizone. Combination of low dose systemic PDE inhibitors might thus be considered for enhancement and in particular prolongation of the lung vasorelaxant response to inhaled iloprost.

P_PA, mean pulmonary artery pressure; P_SA, mean aortic pressure; CO, cardiac output; U46619, stable thromboxane-A₂ analogue; PDE, phosphodiesterase; Mota, motapizone; Sil, sildenafil; Ilo, iloprost.