Anti-Inflammatory and Anti-Apoptotic Activities of Activated Protein C Are Independent of Anticoagulant Activity.

Abstract

Activated protein C (APC), a well known anticoagulant enzyme, reduces mortality in severe sepsis patients and exhibits beneficial effects in multiple animal injury models, including neuroprotective activity in rodent ischemic stroke, attenuation of inflammatory lung injury, and survival benefits in rodent sepsis models. For these in vivo benefits, the relative importance of APC's anticoagulant activity vs. APC's direct cytoprotective effects on cells is unclear. APC anticoagulant activity involves inactivation of factors Va and VIIIa whereas cytoprotection by APC involves two receptors, Endothelial Protein C Receptor (EPCR) and Protease Activated Receptor-1. To distinguish the cytoprotective from the anticoagulant activities of APC, we made recombinant human wild type (wt) APC and a protease domain mutant, 5A-APC (RR229/230AA + KKK191−193AAA), a Gla domain mutant, PTGla-APC (APC with prothrombin residues 1–46)(Smirnov et al JBC 1998) and an active site mutant, S360A-APC. Active site titration and chromogenic assays showed that wtAPC, 5A-APC and PTGla-APC had full enzymatic activity while S360A-APC had none. 5A-APC had almost no anticoagulant activity (< 3 %) whereas PTGla-APC had 300% of wtAPC anticoagulant activity. APC binding to EPCR was assayed as binding to immobilized soluble EPCR and to K293 cells transfected with wtEPCR. We verified that wtAPC and 5A-APC each could bind to EPCR with similar affinity (Kd,app = 37 nM and 29 nM, respectively) whereas, in contrast, PTGla-APC bound very weakly to EPCR (Kd,app > 300 nM ). We then compared the anti-inflammatory and anti-apoptotic activities of the hypo-anticoagulant 5A-APC and the hyper-anticoagulant PTGla-APC to those of wtAPC. To assay APC anti-inflammatory activity, APC-mediated inhibition of LPS-induced TNFα secretion from U937 monocytic cells was determined. The ability of 5A-APC to inhibit LPS-induced TNFα secretion was indistinguishable from that of wtAPC with half-maximum inhibition at 5.4 nM and 6.5 nM, respectively. Neither the enzymatically inactive S360A-APC not the protein C zymogen inhibited LPS-induced TNFα secretion, indicating that a functional APC active site was required. Anti-EPCR antibodies blocking APC binding prevented the anti-inflammatory activity of wtAPC, indicating binding of APC to EPCR on U937 cells was required. The anti-apoptotic activity of each APC species was determined in staurosporine-induced endothelial cell apoptosis assays. Dose-dependent inhibition of apoptosis by 5A-APC was indistinguishable from that by wt-APC with half-maximum inhibition at 0.70 and 2.0 nM, respectively. In contrast to this potent anti-apoptotic activity of wtAPC and 5A-APC, the hyper-anticoagulant PTGla-APC required a 24-fold higher concentration for half-maximal inhibition of endothelial apoptosis. As expected, S360A-APC showed no significant inhibition of endothelial apoptosis. Hence, the 5A-APC variant with < 3% anticoagulant activity exhibited normal anti-inflammatory and anti-apoptotic activities in vitro on monocytic and endothelial cells. These APC variants may be useful for in vivo assessment of the relative importance of APC's anticoagulant vs. cytoprotective activities. In summary, these data highlight important distinctions between structural requirements for APC's anticoagulant functions compared to its anti-inflammatory and anti-apoptotic activities. These structural insights may lead to safer therapeutic APC variants that retain APC's beneficial cytoprotective effects but that have reduced bleeding risk due to reduction in anticoagulant activity.