Abstract
The protein C (PC) pathway plays a major role in the interface between coagulation and inflammation. APC has both anticoagulant and anti-inflammatory properties and is the only effective treatment in patients with severe sepsis. However, assessment of APC’s therapeutic effect on other complex disease models has been compromised by its short half-life (15 min) and by difficulties in monitoring protein levels. To overcome this limitation we used adeno-associated viral (AAV) vectors encoding the PC zymogen or APC for hepatocyte specific gene expression. For direct APC secretion we introduced an extra cleavage site adjacent to the activation peptide for the intracellular protease PACE/furin. Three dose cohorts of C57Bl/6 mice (n=4–6 per group) were injected for either AAV-APC or AAV-PC. A single vector injection resulted in continuous sustained long-term PC or APC expression without signs of liver toxicity. APC functional activity was restricted to AAV-APC-treated mice in which APC plateau levels of 88±43, 162±48, or 263±64 ng/ml were determined in a dose dependent manner. Further, AAV-APC expression consisted mainly of APC because no PC was detected by a zymogen specific ELISA. Only APC expressing mice presented enhanced anticoagulation as determined by 11 to 41 % prolongation of the aPTT values (p<0.05–0.005) and decreased thrombin/antithrombin III complex (TAT) levels (from 30 at baseline to 20, 14, or 12 ng/ml, p<0.05–0.0005). Next, we tested whether APC or PC would provide protection against vascular injury at both micro- and macrocirculation levels of living animals. No thrombus formation was detected in APC expressing mice (n=4) following FeCl3-injury of the carotid artery in contrast to uninjected or PC expressing controls (7 thrombi in 7 mice, p<0.01). Anticoagulant efficacy was then evaluated by real-time imaging of thrombus formation following laser induced arteriole injury using widefield intravital microscopy. In AAV-APC treated mice we observed dose dependent anticoagulation: 8 thrombi /12 injury sites in mice expressing ~80 ng/ml, 3/10 at ~160 ng, and 1/7 at ~260 ng/ml APC compared to 42/42 in untreated controls (P<0.001-0.0001). Expression of PC resulted in prevention of thrombus formation only at the highest expression levels of 4000 ng/ml (5/7, p<0.02) but not at 2000 ng/ml (10/10). When these animals were challenged by tail clipping, blood loss was increased only for mice with the highest APC levels by 2-fold (p<0.05). Moreover, at all levels of APC no changes in wound healing rates were observed following punch biopsy. Treatment of homozygous mice for the factor V Leiden (FVL) mutation with the same vector doses (n=3/group) resulted in a similar anticoagulant effect based on the aPTT with 18–27 % prolongation (p<0.05), or based on TAT-levels, dropping from 56.9 ng/ml at baseline to 28.1, 12.9, or 8.0 ng/ml( p<0.05–0.0005). This data shows that continuous expression of APC can overcome the inherited proteolytic resistance of FVL to APC. In summary, these results demonstrate that APC levels, within the range already obtained in humans by protein infusion (up to 400 ng/ml), provide antithrombotic activity dependent on the injury and/or vessel size. In our model, human APC levels of 160 ng/ml present effective anticoagulant effect without increasing the risk of bleeding. This strategy ensures easy assessment of the role of APC in complex disease models at closely defined circulating levels.
Genotyping Error Detection Through Tightly Linked Markers
