Abstract

Factor Xa (fXa) inhibitors are being tested in the clinic for the prevention and treatment of deep vein thrombosis (DVT) following orthopedic surgery. The antithrombotic efficacy of these drug candidates has traditionally been established in animal models as it is not known whether fXa amidolytic activity, activated partial thromboplastin time (aPTT) or prothrombin time (PT) predict efficacious doses. The present study was designed to test the hypothesis that the potency of fXa inhibitors against fXa incorporated into the prothrombinase complex would predict in vivo antithrombotic efficacy. Eight fXa inhibitors from four structurally distinct chemical series with a range of activities against fXa were tested for their ability to inhibit the prothrombinase complex in human plasma. Thrombin generation and subsequent cleavage of a specific thrombin substrate was used as a measure of prothrombinase activity, inhibitory activity being defined by the concentration of inhibitor required to produce a 2-fold extension in the time to maximal thrombin production (2x lag). In vitro rabbit PTs were also determined. Inhibition in the rabbit DVT model was assessed as previously described (

Thromb Haemost
1994
;
71
:
357
) and related to plasma concentrations of drug.

AgentfXa IC50 (nM)Prothrombinase 2x lag (μM)Plasma concentration in DVT (μM)Thrombosis inhibition (%)Rabbit PT 2x change (μM)
PRT50034 0.5 0.18 0.06 94 7.0 
PRT54681 1.3 0.22 1.14 37 2.7 
PRT54676 0.7 0.24 1.65 47 1.7 
PRT54004 0.4 0.25 1.04 47 1.0 
PRT54456 0.8 0.34 3.39 41 1.5 
PRT56848 4.4 0.92 5.2 11 4.7 
PRT54955 3.5 1.35 4.6 19 8.8 
PRT57106 8.2 1.66 9.2 64 
AgentfXa IC50 (nM)Prothrombinase 2x lag (μM)Plasma concentration in DVT (μM)Thrombosis inhibition (%)Rabbit PT 2x change (μM)
PRT50034 0.5 0.18 0.06 94 7.0 
PRT54681 1.3 0.22 1.14 37 2.7 
PRT54676 0.7 0.24 1.65 47 1.7 
PRT54004 0.4 0.25 1.04 47 1.0 
PRT54456 0.8 0.34 3.39 41 1.5 
PRT56848 4.4 0.92 5.2 11 4.7 
PRT54955 3.5 1.35 4.6 19 8.8 
PRT57106 8.2 1.66 9.2 64 

All compounds inhibited soluble fXa by 50 % at concentrations less than 10 nM. However, the rank order of potencies for inhibition of soluble fXa differed from that required to inhibit the prothrombinase complex. There was also poor correlation between the 2x lag value for prothrombinase inhibition and the concentration required to achieve a 2x change in rabbit PT (r2 = 0.57). Neither the activities of fXa inhibition nor the change in rabbit PT predicted activity in the DVT model. In contrast, compounds could be broadly divided into 3 levels of efficacy for inhibition of in vivo thrombus growth depending on their potency in the in vitro prothrombinase assay. PRT50034 had the lowest 2x lag value of 0.18 μM and was the most potent inhibitor of in vivo thrombosis with 94 % inhibition at a plasma concentration of 65 nM. The second group of compounds, with 2x lag values in the prothrombinase assay ranging from 0.22 to 0.34 μM, inhibited in vivo thrombus formation by 37 to 47 % at plasma concentrations ranging from 1.04 to 3.39 μM. Compounds in the third category were the least potent prothrombinase inhibitors (2x lag values greater than 0.92 μM) and were unable to significantly inhibit in vivo thrombosis even at plasma concentrations of 9.2 μM. These data show that the 2x lag value obtained in the prothrombinase assay, and not inhibition of soluble fXa or extension of rabbit PT, is capable of predicting fXa inhibitor efficacy in the in vivo rabbit DVT model.

Disclosures: Employment at Portola Pharmaceuticals, Genelabs, Ilypsa, Chemocentryx, Genentech, and Rigel.; Stock options in Portola, Ilypsa, and Chemocentryx.; Stock options in Genelabs, Genentech, and Rigel.

Author notes

*

Corresponding author