Modulation of Human Platelet Aggregation by the Phosphodiesterase Type 5 Inhibitor Sildenafil

The aim of this study was to investigate if the phospodiesterase type 5 inhibitor sildenafil inhibits collagen-or ADP-induced human platelet aggregation and bleeding time. To investigate this, two studies were designed. In the first, a single oral dose of sildenafil, 100 mg, was administered to healthy men. Bleeding time was determined and agonist (ADP and collagen)-induced platelet aggregation (ex vivo in platelet rich plasma) was measured 0, 1, and 4 h after application. In the second, a single oral dose of sildenafil, 50 mg, was administered and, in addition to the parameters in the first study, we also determined the platelet aggregation after 24 h and measured the effect of a nitric oxide donor (S-nitroso-N-acetylpenicillamine [SNAP]) in combination to mimic a physiologic nitric oxide release from the endothelium. The bleeding time of 1 h after sildenafil medication (100 mg) was significantly prolonged but recovered toward control values after 4 h, whereas application of sildenafil at a lower dose (50 mg) did not alter the bleeding time. Sildenafil (100 and 50 mg) did not inhibit the ADP-induced aggregation, whereas the collagen-induced aggregation (100 mg) was markedly reduced after 1 h and significantly inhibited 4 h after application. This inhibitory effect was overcome by higher concentrations of collagen. SNAP (0.5 μM) induced an inhibition of platelet aggregation that was potentiated after taking sildenafil (50 mg, 1 and 4 h afterward) and abrogated after 24 h. These data indicates that sildenafil may inhibit collagen-induced platelet aggregation ex vivo. After co-administration of nitric oxide, collagen-and ADP-induced platelet aggregation was significantly inhibited, which may reflect physiologic conditions of an in vivo system.

Sildenafil, used largely in the treatment of erectile dysfunction ^(1-3), is a selective inhibitor of cyclic guanosine monophosphate (cGMP)-specific phosphodiesterase type 5, which is mainly located in vascular smooth muscle cells and platelets. Physiologically, nitric oxide is released during sexual stimulation in the corpus cavernosum ⁽⁴⁾ and increases the concentrations of cGMP by activating the soluble guanylate cyclase ⁽⁵⁾. cGMP initializes a signal cascade (partly mediated by cGMP-dependent protein kinases), which leads to vasorelaxation and a subsequent filling of the corpus cavernosum with blood. Physiologically, especially in the corpus cavernosum, cGMP is degraded by phosphodiesterase type 5. Once cGMP is blocked by a phosphodiesterase inhibitor such as sildenafil, its effects are enhanced and the erection is supported. Phosphodiesterases consist of seven families that are responsible for the degradation of cyclic adenosine monophosphate (cAMP) and cGMP ⁽⁶⁾.

Human platelets have been reported to contain three isoforms of phosphodiesterases (types I, III, and V) ⁽⁷⁾. The activation of human platelets can be inhibited by two intracellular pathways regulated by either cGMP or cAMP. However, nitric oxide causes the activation of cGMP-dependent protein kinases, which prevents the agonist-induced activation of myosin light-chain kinase and protein kinase C and inhibits the agonist-induced calcium mobilization from intracellular stores without any major effect on the ADP-regulated cation channel ⁽⁸⁾. Additionally, cGMP causes an increase of cAMP by inhibition of cAMP phosphodiesterases ⁽⁹⁾. Increased cGMP levels inhibit agonist-induced platelet aggregation. Dipyridamole, an ADP uptake and phosphodiesterase type 5 inhibitor ⁽¹⁰⁾, has been extensively used as an antithrombotic agent in clinical applications ⁽¹¹⁾. Moreover, there are reports that phosphodiesterase type 5 inhibitors inhibit platelet aggregation and adhesion in animal models ^(12,13) and that sildenafil exerts antithrombotic effects in combination with a nitric oxide donor in a rat model ⁽¹⁴⁾. These findings indicate that inhibition of phosphodiesterase type 5 may influence platelet aggregation. There is one report that shows in vitro data of human platelets that have been incubated with sildenafil or sodium nitroprusside or both ⁽¹⁵⁾.

We carried out two studies to reinvestigate if sildenafil altered human platelet aggregation. The aim of our first study was to investigate whether a single oral dose of sildenafil (100 mg) inhibited human platelet aggregation. Blood was taken from healthy men before administration, and then 1 h and 4 h after administration of sildenafil. Ex vivo, platelet aggregation induced by collagen or ADP was measured turbidimetrically in platelet-rich plasma and bleeding time was evaluated.

In our second study we investigated whether a single oral dose of sildenafil (50 mg) inhibited human platelet aggregation. Blood was taken from healthy men before administration, and then 1 h, 4 h, and 24 h after administration of sildenafil. Ex vivo, platelet aggregation induced by collagen or ADP was measured and bleeding time was evaluated. Additionally, the influence of the nitric oxide donor S-nitroso-N-acetylpenicillamine (SNAP) ex vivo in combination with sildenafil-treated platelet-rich plasma was evaluated to mimic a physiologic system in vivo with a constitutive release of nitric oxide from the endothelium.

METHODS

Platelet aggregation

This study was approved by the local ethics committee. Blood was obtained from healthy male volunteers (age range, 23-51 years) who had not taken any type of drug in the previous 10 days. Subjects ate a continental breakfast. Blood was taken from 9.30 a.m. to 10.30 a.m., directly before subjects were given of a single oral dose of sildenafil (50 or 100 mg) and subsequently 1 h and 4 h (and at 24 h after 50 mg) afterward and mixed 1:10 (v/v) with acidic citrate dextrose. Platelet-rich plasma was prepared by centrifugation at 280 g for 10 min at room temperature. The platelets were counted with a platelet analyzer (Baker Instruments, Allentown, PA, U.S.A.). Aliquots of platelet-rich plasma (450 μl) were placed in disposable polystyrene cuvettes (LAbor) and aggregation was measured turbidimetrically ⁽¹⁶⁾ (APACT-Fibrintimer, LAbor, Ahrensburg) at 37°C and constant stirring at 1000 rpm. Platelet-poor plasma was obtained by centrifugation (12,000 g for 8 min) and set as 100% translucence. The measurement of aggregation was performed exactly after the same time schedule (experiments starting 90 min after the blood was sampled) and the same protocol. After 20 min of equilibration at 37°C, the aggregation of platelet-rich protein was induced with collagen (Hormon-Chemie, Munich, Germany) or ADP (Sigma). The nitric oxide donor SNAP (Alexis) was prepared freshly and added to the platelet-rich protein 2 min before the addition of ADP or collagen. Platelet aggregation was recorded for 8 min and the extent of aggregation was evaluated by measuring the maximum height reached by the aggregation curves (APACT software).

Bleeding time

We made three superficial incisions on the forearm (1.5-mm in depth) with a standardized lancet after taking each blood sample (directly before administering sildenafil, and after 1 h, 4 h, and 24 h of administration). The blood oozing from the skin wound was periodically blotted and the time until the bleeding stopped was recorded.

Statistical analysis

All data are expressed as means and SEM of n experiments. The results were analyzed by the nonparametric Wilcoxon test. A p value of < 0.05 was considered statistically significant.

RESULTS

Platelet count and volume

The average platelet count of the probands in whole blood was 225,000 ± 14,500 platelets per microliter. Because of the centrifugation, the average platelet count in platelet-rich protein was 421,000 ± 21,000 platelets per microliter. The mean platelet volume in whole blood was 7.8 ± 0.2 μm³. The mean platelet volume of platelets in platelet-rich protein was 7.2 ± 0.1 μm³ (n = 12). Sildenafil did not have any effects on the parameters mentioned above.

Bleeding time

The mean bleeding time of the six men in our first study (who received 100 mg sildenafil orally) was 97 ± 10 s, which was significantly prolonged 1 h after taking sildenafil (167 ± 16 s). This effect was reversible because the bleeding time returned to normal after 4 h (98 ± 16 s) (n = 6) (Fig. 1). In the second study 50 mg sildenafil did not have any significant effect on bleeding time (n = 6) (Fig. 5).

Fig. 1

Fig. 5

Platelet aggregation

Platelet aggregation was induced by collagen or ADP using different concentrations. The aggregation was always measured exactly 90 min after blood sampling to avoid differences of the aggregation due to an altered status of the platelets resulting from ex vivo conditions. ADP at low concentrations induced a reversible platelet aggregation, whereas higher concentrations led to an irreversible aggregation. ADP and collagen caused a dose-dependent aggregation of human platelets in platelet-rich plasma (Figs. 3, 4, 6, and 7). In both studies (50 and 100 mg) the ADP-induced aggregation was not significantly reduced 4 h after sildenafil administration (Figs. 3 and 6) (n = 4-6).

Fig. 3	Fig. 4	Fig. 6
Fig. 7

In our first study (100 mg sildenafil) collagen (0.2 μg/ml)-induced platelet aggregation was inhibited after 1 h and completely abolished after 4 h by sildenafil in vivo (Figs. 2 and 4) (n = 6). This effect was overcome when higher concentrations of collagen (0.375; 0.75 μg/ml) were used to induce the aggregation, resulting in an aggregation that was higher than 70%. Figure 2 shows original tracings of the collagen (0.2 μg/ml)-induced platelet aggregation and the effect of sildenafil (100 mg). In our second study (50 mg sildenafil) an inhibition of the collagen-and ADP-induced platelet aggregation was shown, although this effect was not significant (n = 3-6) (Figs. 6 and 7).

Fig. 2

Additionally, we added SNAP (0.5 μM) 2 min before ADP-or collagen-induced aggregation. SNAP inhibited the collagen-or ADP-induced platelet aggregation slightly but not significantly (Figs. 8 and 9) (n = 3-6). When SNAP was added to platelet-rich plasma (collagen-induced aggregation) of probands who had taken sildenafil (50 mg) the aggregation was completely abolished (Fig. 9) after 1 h and significantly reduced after 4 h (n = 3-5) (comparison of untreated platelet-rich plasma + SNAP vs. sildenafil-treated platelet-rich plasma + SNAP). Twenty-four hours after sildenafil administration, the aggregation was back to control values (Fig. 9). Moreover, the ADP-induced platelet aggregation was markedly inhibited (50 mg sildenafil) after 1 h and significantly reduced after 4 h of sildenafil administration (n = 6) (Fig. 8). This effect was abrogated after 24 h.

Fig. 8

Fig. 9

Side effects

Sildenafil was generally well tolerated. Common side effects in our studies were flushing, slight headache, slight dizziness, and stuffy nose. One proband experienced temporary visual changes of blue and green. The side effects declined after 2 to 3 h.

DISCUSSION

Sildenafil is a potent inhibitor of the cGMP-specific phosphodiesterase type 5, which is located not only in the corpus cavernosum (the main target of this drug), but also in other cells ⁽¹⁷⁾, such as platelets ^(7,17,18). Because cyclic nucleotides are important inhibitors of platelet activation and aggregation ^(8,19,20), we investigated whether medication with sildenafil in healthy probands altered platelet aggregation in platelet-rich plasma.

First, we investigated the effect of a single oral dose of 100 mg sildenafil. We measured a significantly prolonged bleeding time after 1 h and an inhibition of collagen (submaximal concentration)-induced aggregation after 4 h, and then performed a second study with a slightly different design. In the second study we carried out experiments using 50 mg of sildenafil (average dose) and measured the aggregation after 24 h. Moreover, we used the nitric oxide donor SNAP in our aggregation measurements to imitate a physiologic situation in which nitric oxide is permanently available from the endothelium.

We used ADP and collagen in different concentrations to induce platelet aggregation. We always started our measurement of platelet aggregation 90 min after taking the first blood sample (9.30 a.m. to 10.30 a.m.), following exactly the same time protocol (± 5 min) to avoid differences in platelet aggregation caused by differences in the duration that the platelet rich plasma was ex vivo. There have been findings that the platelet aggregability may change time-dependently when stored at room temperature ^(21,22). We cannot rule out that the time of the sampling alone causes a variation in the platelet aggregation because we did not perform a placebo-controlled study. It is generally thought that the platelet aggregability is increased in the morning compared with the night ^(23,24). During daytime several factors, such as increased activity, may cause intraindividual variations of platelet aggregation.

ADP induced a dose-dependent aggregation that was not significantly inhibited by a sildenafil dose of 100 mg. In contrast, a submaximal aggregation (approximately 50%) induced by collagen (0.2 μg/ml) was markedly inhibited 1 h after administration of sildenafil and 4 h later was completely abolished. This inhibitory effect was overcome by higher concentrations of collagen, probably due to the almost maximal aggregation. Changes in aggregation are best seen while using submaximal aggregation stimuli. This was reflected in one experiment in which an additional proband who had been pretreated with aspirin (125 mg/day) showed the same results, although higher concentrations of collagen and ADP were needed to reach the same aggregation level (data not shown). ADP and collagen act on platelets via different receptors, although the underlying signal transduction is not fully understood ^(25-30). However, we cannot relate the inhibitory effect of sildenafil on collagen-induced aggregation to a specific pathway in the signal cascade. Unlike ADP, it is possible that only collagen may trigger an endogenous nitric oxide release of platelets that is activated during activation or aggregation ^(31-34). This nitric oxide/cGMP metabolism in platelets is probably activated during aggregation to prevent exceeding aggregation and to act as a negative feedback mechanism. Until now only collagen could be shown to activate this mechanism of generating cGMP, which is degraded by phosphodiesterase type 5 ^(31,33,35). This might explain why in vitro, without a nitric oxide source such as the endothelium, an ADP-induced aggregation is not inhibited by pretreatment with sildenafil, in contrast to a collagen-induced aggregation.

Therefore, in our second study with 50 mg of sildenafil, we added SNAP as a nitric oxide donor to imitate a physiologic nitric oxide release. A single oral dose of 50 mg of sildenafil did not significantly alter the collagen and ADP-induced aggregation, although a tendency toward inhibition occurred. When SNAP was added before ADP and collagen administration, a significant inhibition of both agonists could be shown, which was abrogated after 24 h. We used a SNAP concentration of 0.5 μM because this dose caused a slight inhibitory nitric oxide effect but not a complete inhibition of the aggregation.

These results are in partial agreement with data from the manufacturer (Pfizer, New York, NY, U.S.A.) ⁽¹⁵⁾, who demonstrated that sildenafil increases the antiaggregatory activity of sodium nitroprusside. In contrast to their findings, we found a significant inhibition of a collagen (100 mg)-induced submaximal aggregation. One could argue that Wallis et al. ⁽¹⁵⁾ induced nearly maximal aggregations (approximately 70%), which may cover any effects of sildenafil. Additionally, the design of our study was much closer to the physiologic situation because our probands received sildenafil in vivo, whereas Wallis et al. added sildenafil in vitro to platelet-rich plasma. In agreement with our data, Wallis et al found a potentiation of the antiaggregatory effect of the nitric oxide donor sodium nitroprusside by sildenafil, which is reflected by our data with the newer nitric oxide donor SNAP.

These effects have been shown previously with another phosphodiesterase type 5 inhibitor ⁽¹⁹⁾. Addition of zaprinast, which is 10 times less potent ⁽³⁶⁾, to a nitric oxide solution, resulted in a potent increase of the inhibitory effect (ADP induced) of nitric oxide ⁽¹⁹⁾. Taking these data together, it seems necessary to add nitric oxide during an aggregation ex vivo to imitate the endothelial nitric oxide release. This was confirmed in an animal model in which sildenafil enhanced antithrombotic effects of a nitric oxide donor ⁽¹⁴⁾.

Administration of 100 mg of sildenafil resulted in a significantly prolonged bleeding time after 1 h, which returned to normal after 4 h. Fifty milligrams of sildenafil did not cause a prolonged bleeding time. There was a slight discrepancy between a prolonged bleeding time after 1 h and the inhibition of the aggregation ex vivo after 4 h. We are not able to explain this phenomenon yet, but it seems that additional factors during thrombosis play a role in vivo.

These data indicate that it might be necessary to have a closer look at patients with bleeding disorders who took Viagra or patients who took Viagra in combination with other anticoagulants such as aspirin or coumarin, although this study does not indicate that sildenafil may cause bleeding in healthy men. Moreover, there are no published data as to whether the bleeding time of patients treated with coumarin is influenced by sildenafil. In conclusion, this study indicates that sildenafil inhibits collagen-induced platelet aggregation ex vivo. This effect was greater in magnitude when nitric oxide was co-administered to imitate physiologic conditions such as the constitutive nitric oxide release from the endothelium.

Acknowledgment:

The authors would like to thank Dr. Kai Zacharowski (William Harvey Research Institute, London, U.K.) for a critical reading of the manuscript.