Which might explain the low activity of platelet PAI-1 observed in most studies. However, both our own data and those of other investigators have suggested that platelets may possess a mechanism to preserve PAI-1 in the active configuration for longer periods of time. To investigate this hypothesis, it is critical that the method used to isolate PAI-1 from the platelet is able to capture the molecule in its active form and that WY 14643 citations spontaneous inactivation during the preparatory procedure is prevented. Conventional enzymatic assays for PAI-1 activity are inappropriate for this purpose and multicenter 
evaluations have shown that the majority of assays fail to correctly determine the true activity of prepared samples, a conclusion that was confirmed by inconsistent and disparate results in our pilot studies. In agreement with our findings Fay et al showed that the amount of active PAI-1 in a porcine coronary artery thrombi was 36%�C50%. However, this result could not be confirmed in in vitro activated human platelets, although gentle conditions for PAI-1 isolation were used. One reason for this might be that neither tPA was present at the time of platelet activation, nor were any other actions taken to stabilize the active form of PAI-1 which could therefore spontaneously have been inactivated during the long time of extraction. To ensure an immediate capture of active PAI1 at the time of lysis and to circumvent the limitations of enzymatic methods, we used an approach in which tPA was present already when the washed platelets were lysed. By subsequent direct detection of tPA and tPA-PAI-1 complex BIBW2992 439081-18-2 formation with antibodies and 125I-tPA, the intricate interactions of the platelet lysate with the enzymatic assays are avoided. Both detection methods indicated that at least 50�C70% of PAI-1 in washed platelets was present in an active configuration that was biologically functional and could bind tPA. Using a conservative definition of the amount of active PAI-1 by using the tPA concentration immediately below the maximum of complex formation, our approach may even have lead to an underestimation of the true amount of active PAI-1. Also, calculation of the proportion of active PAI-1 is dependent on the PAI-1 antigen assay used. In this study PAI-1 antigen was determined by three different ELISA assays which detect all molecular forms of PAI-1 with similar efficiency. We report the activity concentrations calculated from the assay that measured the highest antigen concentrations to avoid a possible overestimation of the activity level. The ELISA assays are optimised for plasma samples, but the concentration of platelet PAI-1 is in accordance with previous reported levels and variations between the assays are probably due to inter-assay variations previously described. A limitation of the functional assay approach is that it only gives an approximate estimate of the activity, since it is limited by the tPA titration intervals. By decreasing the intervals, a 10% difference in the concentration of active PAI-1 could be detected. To shed light on possible mechanisms behind the low activity rates observed in previous studies, we investigated the influence of commonly used pre-analytic procedures. First, we studied the effect of sonication,since a recentstudy has demonstrated that energy levels as low as 30 W may cause protein damage and it is conceivable that a thermodynamically unstable molecule, such as active PAI-1, is more susceptible to inactivation.