Abstract 1104

BAX 499, an aptamer that binds with a high affinity to TFPI (KD ∼0.1 nM), blocks the anticoagulant action of TFPI in plasma and model systems. BAX 499 was recently tested in a phase 1 safety, tolerability and ex vivo efficacy study in hemophilia patients. The study was prematurely stopped due to an increased number of bleeding events. Baxter subsequently performed several analyses to investigate why clinical safety was inconsistent with preclinical observations.

Analysis of plasma levels of full length TFPI (fl-TFPI) and BAX 499, of global hemostatic parameters via thrombin generation (CAT) and thromboelastography (ROTEM) demonstrated that BAX 499 induced a considerable increase in fl-TFPI plasma levels: over 25-fold at the highest dose (72 mg, SubQ). Pharmacodynamic assessment showed that high fl-TFPI concentrations substantially reduce thrombin generation even at excess BAX 499 plasma concentrations (∼1 μM). This strongly suggests that the BAX 499-induced rise in plasma fl-TFPI levels caused the increased bleeding.

Assessment of biological mechanisms underlying the increase in fl-TFPI plasma levels showed that BAX 499 can potentially raise plasma levels by 1) increasing the synthesis of TFPI by endothelial cells, 2) affecting the distribution of TFPI which is bound to the endothelial surface, stored in platelets and circulated in plasma, 3) interfering with the metabolism of plasma TFPI, i.e. preventing the conversion of full length to truncated TFPI by proteases and 4) affecting TFPI clearance.

BAX 499 treatment of Human Umbilical Vein Endothelial Cells (HUVECs) hardly affected TFPI mRNA production, indicating that BAX 499 has virtually no effect on TFPI synthesis at the transcriptional level.

Treatment of HUVECs with BAX 499 stimulated release of TFPI to the cell culture supernatant in a dose dependent manner. FACS analysis showed that BAX 499 has a minimal influence on cell surface TFPI of endothelial cells but, like Heparin, is able to mobilize and release intracellularly-stored TFPI. TFPI released from HUVECs, however, is too low to explain the increased plasma fl-TFPI levels detected in clinical samples.

Most TFPI circulating in blood is truncated (80–90%) and bound to lipoproteins. The Lys86-Gln90 region in TFPI is a hot spot for proteolytic cleavage by a wide variety of proteases such as elastase. Cleavage results in the simultaneous removal of TFPI Kunitz 1 domain and loss in activity. BAX 499 (1μM) caused a ∼4-fold decrease in elastase-catalyzed cleavage of TFPI, showing that BAX 499 can delay the proteolytic processing of full length TFPI.

BAX 499 was tested in Biacore studies for the interference of fl- TFPI binding to low density lipoprotein receptor-related protein 1 (LRP1), involved in receptor-mediated endocytosis.BAX 499 (1 μM) strongly reduced TFPI binding to LRP1.

Pharmacokinetic studies in mice showed that injected human fl-TFPI is rapidly cleared from their circulation and that administration of TFPI with a molar excess of a non-PEGylated variant of BAX 499 resulted in prolonged TFPI clearance. This effect was even more pronounced when TFPI was dosed with PEGylated BAX 499.

In conclusion, the elevated TFPI plasma levels in hemophilia patients treated with aptamer BAX 499 can be explained as follows: BAX 499 releases intracellularly-stored TFPI, impacts its metabolism and prolongs the circulatory half life of fl-TFPI, most likely due to binding of BAX 499 to the Kunitz 3-C terminus domain of TFPI, a region required for fl-TFPI clearance. The net result of these effects is elevated plasma fl-TFPI, which even at a molar excess of BAX 499 retains anti-coagulant activity.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.