In a recent publication, Marshall et al1 reported that they were unable to confirm our findings2 that cyclic guanosine monophosphate (cGMP) and the cGMP-enhancing drug sidenafil promote platelet activation induced by von Willebrand factor (VWF) and low-dose thrombin. They also contradict our findings3,4 on the role of extracellular signal-responsive kinase (ERK) in platelet activation. We would like to highlight some differences between their studies and ours that may explain this controversy.
(1) Marshall et al question the physiologic significance of the platelet VWF receptor, glycoprotein Ib-IX (GPIb-IX), in mediating signaling, mainly because VWF induces “weaker” signals compared with agonists such as collagen. However, whether a signal pathway is important cannot be determined by the strength of detectable “signals” but only by the outcome of the signals. In fact, it is well documented that GPIb-IX signaling is required for stable platelet adhesion and spreading on VWF under both flow and static conditions, and that it also induces integrin activation, platelet secretion, and aggregation.5-11 Thus, GPIb-IX signaling is physiologically significant.
(2) Marshall et al fail to induce full-scale platelet aggregation with 0.1 U/mL thrombin, which was thus referred to as “subthreshold.” However, in our studies, platelets fully aggregated in response to a much lower dose of thrombin (0.025 U/mL).2 This difference suggests that the platelets used in their study were desensitized. One possible desensitization mechanism is activation of the platelets during their isolation (point no. 3). In our study,2 protein kinase G (PKG) knock-out mouse platelets showed a reduced aggregation response to low-dose thrombin (0.025 U/mL) compared with full-scale aggregation of wild-type platelets. Platelet aggregation in response to high concentrations of thrombin was not significantly affected. Thus, the inability of wild-type platelets to respond to low-dose thrombin may represent a possible cause for the inability of Marshall et al to repeat our finding that PKG knock-out platelets had reduced platelet aggregation in response to low-dose thrombin.
(3) The inability of the platelets prepared by Marshall et al to respond normally to GPIb-IX-dependent agonists is also documented in another recent paper published by the same group,12 in which VWF failed to induce functional responses in washed platelets, in contradiction to other investigators in the field.5,7,11 The different responsiveness of platelets to VWF stimulation in their studies thus provides a possible explanation why Marshall et al failed to show cGMP elevation, ERK phosphorylation, and stimulatory effects of cGMP in VWF-stimulated platelets. It is not clear why platelets prepared by Marshall et al respond differently from platelets prepared by other investigators. However, we note that the basal platelet cGMP level determined by Marshall et al is dramatically higher than that reported by us using the same detection method and by other investigators (Table 1). It is therefore possible that the cGMP pathway was inadvertently activated during platelet preparation in their study, which may explain why no further elevation in cGMP levels was observable in response to VWF. We have observed that the effects of cGMP on platelet activation are biphasic. If cGMP is already elevated, exogenous cGMP or cGMP-enhancing drugs would not further increase platelet activation but would rather exert the secondary phase inhibitory effect.
(4) Marshall et al show that PKG inhibitors failed to inhibit thrombin-induced ERK phosphorylation. This result contradicts not only our data but also data from their companion paper18 (by Gambaryan et al), in which PKG inhibitors attenuated thrombin-induced ERK phosphorylation and VWF caused a “variable” increase in platelet cGMP levels. Despite the differences in data interpretation and conclusions, Gambaryan et al18 clearly show that PKG inhibitors attenuate platelet aggregation induced by ristocetin and thrombin, and inhibit thrombin-induced ERK phosphorylation, which is consistent with our data.
On the roles of cGMP and glycoprotein Ib in platelet activation: a response
We are grateful for the opportunity to reply to the letter from Du et al in regard to our recent publication in Blood.1
Our study was designed to compare the role of Src family kinases, protein kinase G (PKG), and p42/44 mitogen-activated protein (MAP) kinases in platelet aggregation by glycoprotein Ib-IX-V (GPIb-IX-V). The study involved experiments in washed platelets, platelet-rich plasma, and whole blood (both in vitro and in vivo). Our observations support a role for Src kinases in mediating platelet aggregate formation in whole blood on collagen at an intermediate rate of shear. In contrast, all of the experimental data indicate that the role of cyclic guanosine monophosphate (cGMP) is inhibitory, whereas we were unable to find a major role for p42/44 MAP kinases in modulating platelet activation.
In their letter, Du et al have questioned whether the methods used to prepare platelets in these studies have influenced the findings. In this context, we wish to emphasize the internal consistency of the observations in washed platelets, platelet-rich plasma, and whole blood. Furthermore, our study and the accompanying paper by Gambaryan et al2 have shown that the 2 PKG inhibitors, Rp-8-pCPT-cGMPS and Rp-8-Br-PET-cGMPS, are able to inhibit platelet function through a mechanism other than that of PKG blockade.1,2 This was illustrated in our study by the observation that the 2 PKG inhibitors inhibit aggregation to thrombin in PKG-deficient murine platelets.1 This challenges one of the major lines of evidence in support of a role for PKG in mediating platelet activation reported by Li et al.3
Our comments on the individual points raised by Du et al are as follows:
We are in agreement with Du et al in regard to the ability of von Willebrand factor (VWF) to promote platelet activation but believe that this alone is not proof of its physiologic relevance.
The experiments on the PKG-deficient platelets were designed to investigate whether the PKG inhibitors have effects that are unrelated to PKG blockade. This was found to be the case. We have also questioned the argument that thrombin mediates platelet activation through the same pathway as GPIb-IX-V (Marshall et al1 ). In this context, the studies by Li et al3 in PKG-deficient platelets would be of increased relevance if undertaken with VWF.
It is misleading to imply that we are the only group to report the absence of platelet activation by GPIb-IX-V in washed platelets, as illustrated by a recent publication in Blood.4 It is widely recognized that the preparation of VWF can have an important influence in such studies. Further, Du et al have failed to emphasize that our study also reported GPIb-IX-V-mediated platelet activation in platelet-rich plasma.5 This is a more relevant assay than washed platelets in that there is no need to add VWF.
Du et al have understandably questioned the high level of cGMP that was reported in the online prepublished version of our paper.1 This has been corrected in the final published version. Thus, the level of cGMP is similar to that reported by others, including the group of Du et al. Further, we measured a similar level of cGMP in plasma and in washed platelets and observed cGMP elevation in response to a number of agents but not to VWF.
The papers by Marshall et al1 and by Gambaryan et al2 also describe inhibition of platelet activation by the PKG inhibitors, 8-pCPT-cGMPS and Rp-8-Br-PET-cGMPS, in agreement with the observations of Li et al.3 However, these 2 studies attribute the inhibitory action to a non-PKG-dependent action. The third PKG inhibitor used by Li et al,3 KT5823, is unable to inhibit purified PKG.6 Additionally, neither we nor Gambaryan et al observed activation of p42/44 MAP kinases by VWF.1,2 Further, we reported in 1996 that the p42/44 MAP kinase pathway is not required for platelet activation by thrombin,7 in contrast to the later study by Li et al.8
While we recognize that experimental conditions can have an important bearing on results, we do not believe that this is the explanation for the opposing observations/conclusions made by us1 and by Gambaryan et al,2 to those made by Du et al. We emphasize that our conclusions are based on the use of several distinct experimental approaches and include studies performed in whole blood, platelet-rich plasma, and washed conditions.
Correspondence: Yotis Senis, Centre for Cardiovascular Sciences, Division of Medical Sciences, Institute of Biomedical Research, Wolfson Drive, The Medical School, University of Birmingham, Edgbaston, Birmingham B15 2TT, United Kingdom; e-mail: y.senis@bham.ac.uk.
