In a recent review by Nurden (1997), it was suggested that a logical consequence of platelet activation in suspension should be a transient decrease in platelet adhesiveness. Indeed, in the present study we describe a transient adhesion refractoriness of activated platelets in suspension. These results suggest a negative feedback at the site of a growing thrombus in which released agonists from the adherent and activated platelets induce adhesion refractoriness in circulating platelets. This conclusion is based on the following observations: (i) circulating platelets in the CPA system lost their adhesion properties after the 1st run (Fig 1); (ii) this refractoriness induction was dependent on the capability of the platelet to be activated, as both metabolic (by PGE1) and receptor (by RGDS or ReoPro) inhibitors eliminated this response (Fig 2). Further support was given by the lack of refractoriness of Glanzmann's thrombasthenia platelets in this system (Fig 2) and the specificity of the matrix that can induce refractoriness. Thus, only ECM-, VWF- and collagen-coated but not BSA-, fibrinogen- or fibronectin-covered plates could induce this response (Fig 2). (iii) the refractory state of circulating platelets was mediated by agonists released from adherent activated platelets, as pretreatment of WB by apyrase (an ADP scavenger) or ADP receptor antagonist inhibits this response (Fig 3). In addition, suboptimal activation of platelets using either ADP or TRAP induced the same adhesion refractoriness (Fig 3).
The term ‘platelet refractoriness’ has been used to describe more than one phenomenon. Grant et al (1976) and Hagen et al (1977) were the first to describe a refractory state for ristocetin of disaggregating platelets following exposure to an aggregating agent. It was further characterized by Mills et al (1990), who showed that membrane alterations, accompanying shape change caused by an aggregating agent, were responsible for inhibiting subsequent agglutination by VWF. Other studies suggest that platelet activation decreases the binding of VWF to GPIb/IX on platelet membranes (Michelson & Barnard, 1987; George & Torres, 1988). However, in a more recent study, White & Rao (1996) described the refractory state of activated platelets in the presence of calcium ions and demonstrated that the refractory platelets remained sensitive to agglutination by ristocetin, indicating that GPIb/IX receptors are still present on agonist-activated platelets. ADP-induced platelet refractoriness of turbidimetrically measured aggregation has also been reported (Rozenberg & Holmsen, 1968; Holme et al, 1977; Peerschke, 1985). In gel-filtered platelets, the refractoriness was accompanied by shape change induced by low or optimal ADP concentrations. Partial recovery of the aggregability and a return of discoid morphology of the platelets following addition of apyrase were observed. Further studies with ADP demonstrated that, at low concentrations, it initiates platelet shape change without aggregation but reduces platelet adhesion to collagen (Meyer et al, 1981). In all the above studies, except for Meyer et al (1981), platelet refractoriness was induced by optimal concentrations of agonists. This resulted in full platelet activation accompanied by release reaction and formation of irreversible aggregates. The dissociation of these aggregates by PGE1 treatment (White & Rao, 1996) or any other means released platelets that were refractory to a second stimulation by any platelet agonist. In contrast to previous studies, in this study, platelet refractoriness was apparently associated with microaggregate formation, which may occur during a physiological haemostatic process, presumably owing to release of a suboptimal level of agonists. The microaggregates formed under these conditions were transient, as well as the adhesion refractoriness to ECM. The refractory platelets described in this study were probably different from those in most of the previous studies, but similar to those described by Meyer et al (1981), deriving from partial activation and microaggregate formation that spontaneously disaggregate within a period of 10–30 min. During the period that platelets are in microaggregates, they do not adhere to the ECM (or immobilized VWF) under flow conditions. As soon as the microaggregates dissociate, platelets regain their normal properties of adhesion to ECM under flow conditions. The microaggregate formation described in this study is similar to the previously described low ADP concentration-dependent aggregation (Frojmovic et al, 1983) that was distinct from high ADP concentration-dependent aggregations. These two types of aggregates are probably two successive steps in platelet aggregation with shape change and reversible aggregation in the first step, followed by stable aggregation associated with release from dense granules (Holmsen, 1982). It is noteworthy that, in our system, the adhesion refractory state was associated with changes in cell surface receptors that preceded the transient microaggregate formation (Fig 5).
This study suggests a role for platelet agonists at suboptimal concentrations in modulating platelet function and limiting the expansion of the thrombus. The results of this study should, however, be considered cautiously. The CPA system consists of a very small volume of WB that continuously circulates over the ECM for about 120 s in order to achieve a full refractory response. As a consequence of these conditions, the concentration of the released agonists may be increased above physiological levels. Therefore, more studies using flow-through chambers are necessary to elucidate the potential role of such a mechanism in vivo. However, it may be speculated that, in the microcirculation and especially at a bleeding site where a thrombus is being formed, static (or close to static) conditions may be present that allow an accumulation of the released agonists to levels similar to those operating in our assay system (Goldsmith & Karino, 1987). Furthermore, the role of ADP (at optimal concentrations) released by activated platelets in the recruitment and activation of platelets at the site of thrombus formation is well established. It is reasonable to assume that suboptimal concentrations similar to those achieved in our in vitro system may exist downstream of a growing thrombus and, thus, limit its expansion.