Rivaroxaban and Apixaban Induce Clotting Factor Xa Fibrinolytic Activity R

Home , Vitamin K antagonist

Rivaroxaban and apixaban induce clotting factor Xa fibrinolytic activity R. L. R. CARTER ,* † K. T A LB OT ,* †‡ W. S. HUR,* § S. C. MEIXNER,* †‡ J. G. VAN DER G U G T E N, † D. T. HOLMES, † H. C. F. CÔTÉ, * † C. J. KAST RUP , * § T . W . S MI T H , *** A . Y . Y . L EE *** and E. L. G. PRYZDIAL* †‡ *Centre for Blood Research, University of British Columbia; †Department of Pathology and Laboratory Medicine, University of British Columbia, Vancouver, British Columbia; ‡Centre for Innovation, Canadian Blood Services, Ottawa, Ontario; §Michael Smith Laboratories and Department of Biochemistry and Molecular Biology, University of British Columbia, Vancouver, British Columbia; Department of Pathology and Laboratory Medicine, St Paul’s Hospital, Vancouver, British Columbia; and **Division of Hematology, Department of Medicine, University of British Columbia, Vancouver, British Columbia, Canada

This is the peer reviewed version of the following article: [Carter RLR, Talbot K, Hur WS, Meixner SC, Van Der Gugten JG, Holmes DT, Côté HCF, Kastrup CJ, Smith TW, Lee AYY, Pryzdial ELG. Rivaroxaban and apixaban induce clotting factor Xa fibrinolytic activity. J Thromb Haemost 2018; 16: 2276–88.], which has been published in final form at [https://onlinelibrary.wiley.com/doi/full/10.1111/jth.14281]. This article may be used for non-commercial purposes in accordance with Wiley Terms and Conditions for Use of Self- Archived Versions.

Essentials: • Activated clotting factor X (FXa) acquires fibrinolytic cofactor function after cleavage by plasmin. • FXa-mediated plasma fibrinolysis is enabled by active site modification blocking a second cleavage. • FXa-directed oral anticoagulants (DOACs) alter FXa cleavage by plasmin. • DOACs enhance FX-dependent fibrinolysis and plasmin generation by tissue plasminogen activator.

Keywords: direct oral anticoagulant; factor Xa; fibrinolysis; plasmin; tissue-type plasminogen activator.

Abstract

Background: When bound to an anionic phospholipid-containing membrane, activated clotting factor X (FXa) is sequentially cleaved by plasmin from the intact form, FXa, to FXa and then to Xa33/13. Tissue-type plasminogen activator (t-PA) produces plasmin and is the initiator of fibrinolysis. Both FXa and Xa33/13 enhance t-PA-mediated plasminogen activation. Although stable in experiments using purified proteins, Xa33/13 rapidly loses t- PA cofactor function in plasma. Bypassing this inhibition, covalent modification of the FXa active site prevents Xa33/13 formation by plasmin, and the persistent FXa enhances plasma fibrinolysis. As the direct oral anticoagulants (DOACs) rivaroxaban and apixaban bind to the FXa active site, we hypothesized that they similarly modulate FXa fibrinolytic function. Methods: DOAC effects on fibrinolysis and the t-PA cofactor function of FXa were studied in patient plasma, normal pooled plasma and purified protein experiments by the use of light scattering, chromogenic assays, and immunoblots. Results: The plasma of patients taking rivaroxaban showed enhanced fibrinolysis correlating with FXa. In normal pooled plasma, the addition of rivaroxaban or apixaban also shortened fibrinolysis times. This was related to the cleavage product, FXa, which increased plasmin production by t-PA. It was confirmed that these results were not caused by DOACs affecting activated FXIII- mediated fibrin crosslinking, clot ultrastructure and thrombin-activatable fibrinolysis inhibitor activation in plasma. Conclusion: The current study suggests a previously unknown effect of DOACs on FXa in addition to their well-documented anticoagulant role. By enabling the t-PA cofactor function of FXa in plasma, DOACs also enhance fibrinolysis. This effect may broaden their therapeutic indications.

Introduction

Positioned at the convergence of the intrinsic and extrinsic branches of coagulation, the activated form of factor X, FXa, is an important target for drug development [1–3]. FXa is localized to the site of vascular damage by calcium-dependent association with the cofactors activated FV and anionic phospholipid-containing membrane (PL), where it converts prothrombin to thrombin. Thrombin is multifunctional. It catalyzes the conversion of soluble fibrinogen to a fibrin clot, and feedback amplifies coagulation [4]. The clot is also protected by thrombin from premature fibrinolysis by: (i) activation of FXIII, which crosslink- stabilizes the clot [5]; (ii) activation of the carboxypeptidase thrombin-activatable fibrinolysis inhibitor (TAFI) [4,6], which excises C-terminal lysine residues that would normally act as cofactors to enhance fibrinolysis; and (iii) effects on microfibril density and the thickness of fibrin polymers [7], which, at higher thrombin concentrations, adopt an architecture that is resistant to dissolution. These feedback roles played by thrombin in plasma must be considered when the contributions of upstream proteins, such as FXa, are evaluated, as varying conditions may have effects on the amount of thrombin produced during an experiment.

Fibrin is a cofactor for tissue-type plasminogen activator (t-PA), the fibrinolysis initiator [8], thus localizing its own degradation. t-PA activates plasminogen (Pg) to plasmin, which cleaves fibrin and solubilizes the clot. Unlike most proteases in plasma, t-PA circulates as a functional enzyme, but it requires a protein cofactor to exceed the antifibrinolytic threshold of plasma. To initiate plasmin generation, fibrin has weak inherent binding sites for t-PA and Pg, thus improving the enzyme–substrate encounter. This first phase of plasmin production amplifies the cofactor activity of fibrin by exposing C-terminal lysines that are integral to new t-PA-binding and Pg-binding sites. However, at the supraphysiological concentrations used therapeutically and for most published studies, t-PA may produce plasmin systemically, because it has inherent proteolytic activity. This bypasses the first phase of cofactor activity provided by fibrin to produce the initial plasmin that accelerates fibrinolysis [9]. Given these artificially high t-PA concentrations, it is understandable that the prevailing fibrinolysis model accepts fibrin as the only required t-PA cofactor [10–12]. However, a role for other cofactors that may participate in the initial production of plasmin has been under- investigated. Recently, the use of lower t-PA concentrations within the physiological range revealed that FXa and other clotting factors act as auxiliary t-PA cofactors [2,9,13–17]. These are cleaved by plasmin more effectively than fibrin, providing early C-terminal lysine- mediated t-PA enhancement in the context of clotting.

The fibrinolysis cofactor activity of FXa is attained by plasmin treatment or, less efficiently, by autoproteolysis, which first releases a 3-kDa peptide from the C-terminus of intact FXa (FXa), giving FXa [18]. FXa and FXa are discernible by denaturing Laemmle electrophoresis, but have identical clotting activity [18]. However, at least two forms of functionally distinct FXa are not distinguishable by electrophoresis, and their formation depends on calcium-mediated binding of FXa to PL. FXa that binds to Pg and enhances t- PA activity is formed when the starting FXa is bound to PL (Fig. 1A). Under these conditions, a second plasmin-mediated cleavage of FXa yields Xa33/13 [17], which is also fibrinolytic. In contrast, in the absence of either PL or calcium, the FXa and subsequent Xa40 that are produced do not bind Pg (Fig. 1B). Unlike FXa, Xa33/13 and Xa40 lack coagulation activity [18]. The fibrinolytic differences in FXa subpopulations may be explained by the differential exposure of C-terminal lysine versus arginine, which precludes electrophoretic detection, owing to proximity.

When evaluated by the use of purified proteins, Xa33/13 enhances t-PA activity more effectively than the precursor FXa [9]. However, as compared with FXa, the t-PA cofactor function of Xa33/13 is rapidly degraded in plasma [2]. We have reported that irreversible modification of the active site with chloromethyl ketones (CMKs) prevents Xa33/13 production (Fig. 1C), allowing the persistence of t-PA cofactor function in plasma [2]. The current study addresses the hypothesis that, like CMK-modified FXa [2], direct oral anticoagulants (DOACs) specific for the FXa active site (rivaroxaban and apixaban) enhance fibrinolysis by stabilizing FXa. Interestingly, previous studies have investigated the effects of DOACs on fibrinolysis [19–21], and have reported enhanced clot degradation, although differences between thrombin-specific and FXa-specific DOACs were noted. These combined studies support a model in which reduced thrombin production in plasma caused by DOACs attenuated the generation of activated TAFI (TAFIa), making clots more susceptible to fibrinolysis [21]. In these reports, a direct role for FXa in fibrinolysis was not investigated. In the current study, rivaroxaban and apixaban were shown to enhance fibrinolysis by FXa separately from effects on downstream thrombin generation.

Materials and Methods

Reagents and proteins

HEPES, polyethylene glycol 8000 and anhydrous dimethylsulfoxide (DMSO) were from Sigma-Aldrich (St Louis, MO, USA). Tetrasodium EDTA was from Thermo Fisher Scientific (Mullica Hill, NJ, USA). Calcium chloride dihydrate was from EMD Chemicals (Etobicoke, ON, Canada). Innovin was from Siemens Healthcare Diagnostic (Oakville, ON, Canada). Human Lys-Pg was from Enzyme Research Laboratories (South Bend, IN, USA). Tween-20 was from Amresco (Solon, OH, USA). Single-chain recombinant t-PA was from Genetech, South San Fransisco, CA, USA). Plasmin and thrombin fluorogenic substrates (Boc-Glu-Lys-Lys-MCA and Boc-Val-Pro-Arg-MCA, respectively) were from Peptide Institute (Osaka, Japan). The chromogenic substrates H-D-Val-Leu-Lys-pNA (S-2251), Z-D-Arg-Gly-Arg-pNA (S-2765) and H- D-Phe-Pip-Arg-pNA (S-2238) were from Chromogenix (Bedford, MA, USA). L-a- phosphatidylcholine and L-a-phosphatidylserine (PCPS) (Avanti Polar Lipids, Alabaster, AL, USA) at a 3 : 1 ratio (PCPS) were used to prepare small unilamellar vesicles [17]. Human FX, potato tuber carboxypeptidase inhibitor (PTCI), a-thrombin and plasmin were from Haematologic Technologies (Essex Junction, VT, USA). Normal pooled plasma and FX- deficient plasma (FXDP) were from Affinity Biologicals (Hamilton, ON, Canada). Primary polyclonal rabbit anti-human fibrinogen antibody was from Dako (Santa Clara, CA, USA). Secondary horseradish peroxidase (HRP)-conjugated goat-anti-rabbit IgG H&L antibody was from Abcam (Toronto, ON, Canada). Primary mouse anti-human FX heavy chain antibody was from Green Mountain Antibodies (Burlington, NJ, USA). Secondary HRP-congugated goat anti-mouse IgG H&L antibody was from Jackson ImmunoResearch (Westgrove, PA, USA).

Rivaroxaban and apixaban preparation

Apixaban (Eliquis; Bristol-Myers Squibb, Montreal, QC, Canada) and rivaroxaban (Xarelto; Bayer, Mississauga, ON, Canada) tablets were crushed and dissolved in DMSO to obtain 10 mM stock solutions, and centrifuged at 10 000 x g for 5 min to remove insoluble constituents. An FXa chromogenic activity assay (substrate; S-2765) was conducted to verify DOAC specific activity. Prior to use, the stock DOACs were diluted in HEPES-buffered saline (HBS) containing 0.1% polyethylene glycol. The final concentration of DMSO was the same in all experiments, including controls (0.1%), and is 10-fold below the concentration found to affect fibrinolysis (> 1.0%). DOAC concentrations were based on standard therapeutic doses. For rivaroxaban, 0.57 M (250 ng mL-1), which is the maximum plasma concentration (Cmax) reported for patients taking 20 mg daily, was used [21]. For apixaban, the Cmax is, by convention, 0.44 M (200 ng mL-1) [22–24], which is double the Cmax reported after the first of two daily 5-mg administrations [25]. Higher apixaban concentrations have been noted [19,26].

Patient samples

Blood was drawn from consenting patients (University of British Columbia and Canadian Blood Services Research Ethics Boards H11-00084) being treated for atrial fibrillation, which, as compared with other DOAC indications, has minimal effects on hemostasis. This was secondary use of data, and biospecimens were collected for another protocol (University of British Columbia H11-01358). Each patient’s rivaroxaban or dabigatran (a thrombin-selective DOAC) dose and the time of the last dose were recorded. Blood was collected into 0.109 M sodium citrate tubes (BD Biosciences, San Jose, CA, USA), and platelet-poor plasma was stored in aliquots at -80 °C.

Plasma sample concentrations of rivaroxaban and dabigatran were quantified by liquid chromatography and tandem mass spectrometry. Briefly, plasma (50 L) was mixed with internal standard (rivaroxaban-d4/dabigatran-d3; 25 L), and this was followed by base hydrolysis with 0.2 M NaOH (25 L) and protein precipitation with 35 mM ZnSO4 (400 L) in a 96-well plate. Supernatant (10 L) was injected onto a C18 HPLC column (LC-20AD, Shimadzu, Chicago, IL, USA), eluted with a 2 mM ammonium acetate/0.1% formic acid/methanol gradient, and analyzed on a SCIEX 5500 QTrap mass spectrometer (Concord, ON, Canada).

Plasma clot lysis

Turbidity assays were similar to those previously described [15]. Here, DOAC or vehicle (including DMSO), t-PA in HBS/0.01% Tween (HBST) and a mixture containing thrombin (10 nM), CaCl2 (20 mM), and penicillin/streptomycin (1%; Thermofisher Scientific) were added as separate droplets to a 96-well flat-bottomed plate [27]. Unless described otherwise, final concentrations of 0.57 M rivaroxaban, 0.44 M apixaban and 30 pM t-PA were used during plasma clot lysis assays. The assay was initiated with normal pooled plasma, or FXDP ± purified human FX (170 nM). The final plasma concentration was 35%. The extent of clot formation and dissolution was monitored at 405 nm with 5-min intervals. To calculate the area under the curve (absorbance at 405 nm x time) with GRAPHPAD PRISM (GraphPad Software, San Diego, CA, USA), an endpoint was chosen at or near completion of clot lysis for the experimental conditions being evaluated, and was consistently used for the respective controls conducted in the absence of DOACs. Individual sets of data were conducted in triplicate (n = 3) or quadruplicate (n = 4) throughout this study and reported as averages.

Plasmin or thrombin activity during fibrinolysis

Cleavage of a plasmin-selective or thrombin-selective fluorogenic substrate was monitored during plasma clot lysis assays [28,29]. The reactant solution (15 L of DOAC or buffer, t-PA, thrombin, calcium, and penicillin/streptomycin) was added to a black polystyrene 96-well flat-bottomed plate (Corning, Corning, NY, USA), followed by 5 L of the plasmin-selective or thrombin-selective fluorogenic substrate (Boc-Glu-Lys-Lys-MCA and Boc-Val-Pro-Arg- MCA, respectively). After 3 min, 80 L of prewarmed plasma was added, and plasmin generation was monitored on a POLARstar OPTIMA plate reader (BMG Labtech, Ortenberg, Germany) for 3 min at 37 °C by fluorescence (excitation, 390 nm; emission, 460 nm; 30-s intervals). A standard curve obtained with purified plasmin was produced to derive percentage change.

Fibrin Crosslinking

Clots (40 L) were formed as in the plasma lysis assays, but without t-PA. To each sample, 100 L of quench solution (50 mM dithiothreitol, 12.5 mM EDTA, 8 M urea) was added, and samples were incubated at 60 °C until the clots dissolved. Reduced samples were subjected to 10% polyacrylamide Laemmle electrophoresis, transferred to nitrocellulose membranes, and probed for fibrinogen. After three washes (phosphate-buffered saline/0.01% Tween), secondary antibody was applied, and this was followed by additional washes. Bands were visualized with Clarity Western ECL substrate (Bio-Rad, Rutherford, NJ, USA), and the luminescence was detected on X-ray film (Mandel, Guelph, ON, Canada).

Scanning electron microscopy

Plasma clots (100 L) were formed in 12 x 75-mm polypropylene tubes as above. After 1 h, clots were washed, and fixed in a mixture of 2.5% glutaraldehyde and 4% formaldehyde, dehydrated with increasing concentrations of ethanol (30–100%), and critical-point-dried. Platinum/palladium-coated clots were imaged with a Hitachi (Toronto, ON, Canada) S-4700 field emission scanning electron microscope. For each condition, the diameter of 10 non- superimposed fibers was measured at three constant locations on an overlaid grid with IMAGEJ software. Clot porosity was estimated by converting gray-scale images to a binary format and quantifying black pixels (pores).

FXa fragmentation in plasma

Innovin (8 L), CaCl2 (5 mM; 1.5 L), t-PA in 0.01% HBST (50 nM; 1.5 L) and HBS (9 L) were incubated with or without rivaroxaban or apixaban. The reaction was initiated with plasma (10 L) in a total reaction volume of 30 L. Separate reactions were evaluated for each time point. To stop the reaction, 10 L of SDS-Laemmle sample buffer was added for electrophoresis and western blot analysis. Western blots were probed with mouse anti- human FX heavy chain antibody, as previously described [9]. Molecular weight markers (FroggaBio, Toronto, ON, Canada) and purified XaAT, FX, FXa and Xa33/13 were used to identify bands.

Plasmin-mediated fragmentation of FXa

Purified FXa (2 M), PCPS (300 M, and CaCl2 (5 mM) were combined with rivaroxaban or apixaban (FXa/DOAC ratios of 1 : 5.7 and 1 : 4.4, respectively). As above, rivaroxaban and apixaban final concentrations were 11.4 M and 8.8 M, respectively, with 0.2 M plasmin. At the indicated times, samples were withdrawn and added to Laemmle electrophoresis sample buffer for anti-FX western blot analysis, as previously described [9]. t-PA-mediated plasmin generation

The effect of FXa ± rivaroxaban or apixaban on t-PA mediated plasmin generation was measured by use of the plasmin-selective chromogenic substrate S2251 (200 M), as described previously [9]. Samples from the reaction mixture of an identical experiment were removed for anti-FX western blot analysis and [125I]Pg ligand blot analyses, as previously described [9,13].

Results

FXa persistence correlates with enhanced fibrinolysis in rivaroxaban-treated patient plasma

To further investigate the prior correlation between FXa stabilization and enhanced fibrinolysis observed in vitro [2], plasma samples from patients taking DOACs were evaluated. Figure 2A shows that conversion of FXa to Xa33/13 in normal pooled plasma is blocked by the addition of the FXa-specific DOACs rivaroxaban and apixaban, like modification of the FXa active site by CMK. In the plasma of patients treated with the direct thrombin inhibitor dabigatran, the degradation of FXa to Xa33/ 13 was clear (Fig. 2B). However, consistent with pooled normal plasma supplemented with FXa-specific DOACs (Fig. 2A), the production of Xa33/13 in patient plasma treated with rivaroxaban was blocked (Fig. 2C). To evaluate the effect of persistent FXa on fibrinolysis, rivaroxaban-treated patient plasma was compared with normal pooled plasma controls. Clot formation in plasma supplemented with t-PA was induced with thrombin and calcium, and fibrinolysis was monitored by changes in turbidity. Figure 2D shows that the time needed to achieve half clot lysis was markedly reduced in the rivaroxaban cohort. Quantification of FXa and FXa by western blot densitometry revealed an inverse linear correlation to the proportion of FXa and direct linear correlation to the proportion of FXa (Fig. 2E). There was no relationship between fibrinolysis and the circulating concentration of DOAC in patient plasma, time to last DOAC dose, patient age or sex, or time to maximal turbidity under the conditions of the fibrinolysis experiment (data not shown).

DOACs enhance normal pooled plasma clot lysis

Clot formation and fibrinolysis in normal pooled plasma was followed as shown in Fig. 2 to investigate the dose dependence of DOACs. When rivaroxaban and apixaban were added at their respective Cmax values, fibrinolysis was enhanced (Fig. 3A). Reducing the concentration of rivaroxaban and apixaban to below the Cmax increased the length of time required for fibrinolysis (Fig. 3B). This indicated that the concentration of circulating rivaroxaban in patient plasma (Fig. 2) was sufficient to saturate the related mechanism. To confirm that initiation of clot formation by the addition of purified thrombin was sufficient to outweigh the possibility of thrombin generation during the plasma fibrinolysis experiments, thrombin activity was measured directly by use of a fluorogenic substrate in the presence and absence of various combinations of FXa and DOAC. Figure 3C shows thrombin activity during the clot-generation phase of the fibrinolysis experiments shown in Fig. 3A. As expected, when purified FXa was added in the absence of DOAC (+ FXa), there was increased thrombin activity as compared with addition of only thrombin (– t-PA, – cofactor), or any combination of DOAC in the presence or absence of purified FXa. These fluorogenic data show that DOACs did not affect overall thrombin activity, and are in support of the observed effects on fibrinolysis under the current conditions being attributable to FXa–DOAC interactions.

Inhibition of TAFIa does not affect enhanced fibrinolysis by DOACs

Most anticoagulants have been shown to enhance the rate of fibrinolysis through mechanisms involving reduced activation of TAFI because of restricted thrombin production. Previous work has demonstrated that 2 M PTCI completely inhibits the activity of TAFIa during thrombin-induced clot formation [30]. In the current study, we showed that including 2 M PTCI in the reaction mixture had no effect on early fibrinolysis in the presence of rivaroxaban and apixaban when clotting was initiated with 10 nM thrombin (Fig. 4). Enhancement was also seen when the PCTI level was increased to 10 M (data not shown).

Fibrin crosslinking is not affected by DOACs

To further exclude the possibility of thrombin-mediated effects in fibrinolysis of FXa- directed DOACs, western blotting was conducted to examine fibrin generation and crosslinking. The addition of rivaroxaban or apixaban to either normal plasma or FXDP made no difference to the conversion of fibrinogen to fibrin initiated by purified thrombin (Fig. 5). This analysis included the extent of covalent - chain to -chain bond formation attributable to FXIIIa activity.

Fibrin fiber diameter and porosity are not altered by DOACs

Thrombin was further ruled out as a potential variable in our experiments by analysis of clot ultrastructure. Clots were formed in normal plasma as in Fig. 3, with the exclusion of t-PA to minimize dissolution, and analyzed by scanning electron microscopy. Representative electron micrographs are shown in Fig. 6A; these provide further evidence that there are no significant qualitative differences in fibrin structure that account for the effects of rivaroxaban and apixaban on enhanced fibrinolysis. By use of a grid system, the diameter (Fig. 6B) and porosity (Fig. 6C) of fibers were quantified. No significant fibrin ultrastructural differences were revealed, confirming that this was not contributing to enhanced fibrinolysis. A similar result was observed in clots induced with 30 nM thrombin in the presence or absence of DOAC (data not shown).

FX is required for DOACs to enhance plasma clot lysis and plasmin generation

Having eliminated confounding roles for TAFI and fibrin structure in the effects on fibrinolysis of rivaroxaban and apixaban, we established the involvement of FX by using FXDP. Clot dissolution was followed with or without addition of purified FX at the physiological concentration (0.17 M), confirming that FX was required for the fibrinolytic effects of DOACs (Fig. 7A). Notably, when FXDP is used, premature fibrinolysis has been reported [30], and this was seen here. To overcome the anticoagulant effects of DOACs on thrombin generation during the experiment, and consequently on fibrinolysis, other studies have adjusted the initiating tissue factor concentration to equalize clotting times between samples [31,32]. With a similar strategy to compare the fibrinolysis of FXDP with or without purified FX, t-PA was titrated. Quantification of fibrinolysis showed that, at concentrations of t-PA below 25 pM, rivaroxaban and apixaban enhanced fibrinolysis of FXDP, but only when FX was available (Fig. 7B). The effect of FX was titratable to 0.3 M (data not shown).

To further understand the mechanism by which DOACs enhance fibrinolysis, a specific fluorogenic substrate was used to measure the amount of plasmin generated during fibrinolysis in FXDP with or without FX, and in normal plasma. The percentage difference in plasmin between the conditions was derived for comparison. Accounting for the enhanced fibrinolysis, the results showed that plasmin generation was increased in an FX-dependent manner (Fig. 7C).

Plasmin generation in the presence of DOACs correlates with altered FXa fragmentation

Experiments with purified t-PA, FXa and Pg were conducted to exclude the possible contributions of ancillary proteins in plasma. The western blot shown in Fig. 8A shows that Xa33/13 is generated in the absence of DOACs, confirming previous findings [9] and supporting those obtained with plasma in the current study (Fig. 2). The plasma experiments were also consistent with the finding that rivaroxaban or apixaban blocked the conversion of FXa to Xa33/13 when purified proteins were used, with evidence that FXa was, instead, being cleaved to Xa40. In addition to inhibition of Xa33/13 generation, a delay in the conversion of the starting FXa to FXa was clearly shown at 2 min in the presence of DOACs. The corresponding ligand blot (Fig. 8A) showed preferential binding of [125I]Pg to the 33-kDa component of Xa33/13 as compared with FXa, as previously observed [2,9,13–17]. In the presence of DOACs, however, [125I]Pg associated prevalently with FXa, as Xa33/13 was not produced. Newly generated FXa was required for the optimal association with [125I]Pg, which subsequently diminished, as was seen with Xa33/13. This lost function in the presence of DOACs correlated with the appearance of Xa40. Demonstrating specificity (data not shown), the interaction of [125I]Pg with FXa fragments was completely inhibited by the C-terminal lysine analog -aminocaproic acid, as described previously [2,9,13–17]. The same fragmentation profile was obtained when purified FXa was treated with plasmin in the presence of DOACs (Fig. 8B). Thus, FXa-directed DOACs, like CMKs, alter the cleavage of FXa by plasmin.

CMK-modified FXa is known to enhance plasmin generation in a purified system to an extent that is intermediate between that seen with no addition and that seen with purified FXa [9]. Consistent with this finding, rivaroxaban-treated or apixaban-treated FXa increased t-PA-mediated plasmin generation to a level that was intermediate between that seen with no addition and that seen with FXa in the absence of DOAC (Fig. 8C). The explanation for the greatest enhancement by FXa without DOACs is that Xa33/13 production is enabled, whereas this is blocked in the presence of DOACs (Fig. 8A). Insignificant plasmin generation occurred in the absence of t-PA (data not shown).

Discussion

Using patient plasma, normal plasma, and purified proteins, we have shown that rivaroxaban and apixaban enhance fibrinolysis in plasma by preventing the cleavage of FXa to Xa33/13 (Fig. 1). Although Xa33/13 has superior t-PA cofactor function to that of FXa in experiments using purified proteins, this Xa33/13 activity is rapidly lost in plasma [2]. The consequent accumulation of FXa in pooled normal plasma supplemented with rivaroxaban or apixaban results in persistent FXa-derived fibrinolytic activity. Thus, in plasma from atrial fibrillation patients being treated with therapeutic rivaroxaban, there was a direct correlation between enhanced fibrinolysis and the conversion of FXa to FXa. Anticoagulants have been previously shown to accelerate clot dissolution by reducing thrombin production, thereby attenuating the related feedback mechanisms leading to TAFI activation and resistance to fibrinolysis [19–21,33].

These variables were circumvented in the current study by overwhelming the thrombin produced during plasma experiments with excess purified thrombin. Here, the involvement of TAFIa was excluded by use of the potent inhibitor PTCI [19,34]. Similarly, DOACs had no detectable effect on fibrin ultrastructure, including the extent of crosslinking between fibrin subunits, fibrin - dimer formation, fiber diameter, and relative porosity. The uniformity of fibrin crosslinking in the presence or absence of DOACs was corroborated by earlier work showing that FXIII activation was unaffected [21], because sufficient thrombin was generated regardless of the presence of DOACs to facilitate its activation [35]. We also confirmed, by direct thrombin activity measurements, that thrombin produced during the experiment was insignificant as compared with the purified thrombin used to initiate clot formation. Corroborating the direct role of FXa in rivaroxaban-enhanced or apixaban-enhanced plasmin generation by t-PA, fibrinolysis of FXDP was unaffected by DOACs unless purified FX was added. The fact that fibrinolysis of FXDP was hastened even in the absence of DOACs has been reported previously [30–32], and we are investigating the basis of this. The increased rate of fibrinolysis in FXDP may additionally involve attenuation of an FXa-inhibitory effect on fibrinolysis, although experiments using purified proteins contradict this possibility [9]. Nevertheless, the high specificity of rivaroxaban and apixaban for FXa, and the exclusion of thrombin feedback, suggest a role for these DOACs in FXa-dependent enhancement of fibrinolysis. Although anticoagulation by FXa-specific and thrombin-specific DOACs is ultimately restricted by diminished thrombin activity, there are noteworthy clinical distinctions (reviewed in [36]). As compared with the traditional anticoagulant used in atrial fibrillation, i.e. warfarin, the frequency of bleeding events and the incidence of stroke are reduced by apixaban, but not by rivaroxaban[37,38]. The apparent difference between these is surprising, as they target nearly identical positions within the FXa catalytic site [39,40], but it may be attributable to the five-fold higher affinity of apixaban. Interestingly, as compared with the use of warfarin for stroke prevention in patients with atrial fibrillation, apixaban may give a somewhat improved outcome relative to thrombin-specific dabigatran [36,41]. Analogous thrombin-specific DOACs, i.e. argatroban and dabigatran, had no effect on clot dissolution (data not shown) or FXa fragmentation in our studies, possibly explaining how apixaban may reduce comparative risk.

Thus, the enhanced fibrinolysis that is presumably localized by FXa binding to PL (Fig. 1) [9] or directly to fibrin [42] may support the greater safety of apixaban than of dabigatran. This hypothesis does not explain why, in a cross-population post hoc analysis, rivaroxaban caused a higher incidence of major bleeding than did dabigatran when they were compared with warfarin in atrial fibrillation treatment [36,43], supporting the need for further investigation, as previously suggested [44]. The finding that rivaroxaban and apixaban have a role in fibrinolysis may suggest new clinical applications, such as for post-thrombotic syndrome (PTS) [45,46]. A potential basis for PTS is ineffective clot dissolution with residual vascular scarring [47], leading to localized chronic inflammation [48]. Although additional studies are required to determine the effects of DOACs versus heparin or warfarin in attenuating PTS, a small population analysis indicated an association between better clot resolution and reduced PTS [49]. The results presented here suggest that the dual anticoagulant and novel fibrinolytic roles of rivaroxaban and apixaban may underlie this finding, and indicate an added benefit of FXa- specific DOACs.

Addendum

R. L. R. Carter designed, analyzed and conducted experiments, and wrote the manuscript. K. Talbot designed and conducted preliminary experiments. W. S. Hur conducted experiments and revised the manuscript. S. C. Meixner assisted with experiments and revised the manuscript. J. G. van der Gugten quantified rivaroxaban in patient plasma. D. T. Holmes directed the rivaroxaban quantification and revised the manuscript. H. C. F. Côté assisted in patient data analysis, and revised the manuscript. C. J. Kastrup provided intellectual content and revised the manuscript. T. Smith recruited patients, provided intellectual content, and revised the manuscript. A. Y. Lee provided intellectual content and revised the manuscript. E. L. G. Pryzdial designed and analyzed experiments, provided project direction, and wrote the manuscript.

Acknowledgements

This research was funded by the Heart and Stroke Foundation of Canada (E. L. G. Pryzdial; G- 13-0001644), by a University of British Columbia graduate scholarship (R. L. R. Carter) and a Canadian Institutes of Health Research grant (C. J. Kastrup; FDN-148370). We acknowledge D. Horne (University of British Columbia, Imaging Facility, Faculty of Medicine) for electron microscopy training, and C. Overall (University of British Columbia, Centre for Blood Research, Faculty of Dentistry) for access to a PolarStar fluorometer.

Disclosure of Conflicts of Interest

A. Y. Lee has served on a steering committee at Bayer, and reports receiving grants and study drugs from Bristol Myers Squibb during the conduct of the study. E. Pryzdial has patent 9579367 issued. C. Kastrup reports receiving grants from the Canadian Foundation for Innovation, the BC Knowledge Development Fund, and the Michael Smith Foundation for Health Research during the conduct of the study, as well as grants from Novo Nordisk and personal fees from CoMotion Drug Delivery Systems outside the submitted work. The other authors state that they have no conflict of interest.

References

1 Turpie AG. Oral, direct factor Xa inhibitors in development for the prevention and treatment of thromboembolic diseases. Arterioscler Thromb Vasc Biol 2007; 27: 1238–47. 2 Pryzdial EL, Meixner SC, Talbot K, Eltringham-Smith LJ, Baylis JR, Lee FM, Kastrup CJ, Sheffield WP. Thrombolysis by chemically modified coagulation factor Xa. J Thromb Haemost 2016; 14: 1844–54. 3 Camire RM. Bioengineering factor Xa to treat bleeding. Thromb Res 2016; 141(Suppl. 2): S31–3. 4 Von dem Borne PA, Bajzar L, Meijers JCM, Nesheim ME, Bouma BN. Thrombin-mediated activation of factor XI results in a thrombin-activatable fibrinolysis inhibitor-dependent inhibition of fibrinolysis. J Clin Invest 1997; 99: 2323–7. 5 Lor and L. Factor XIII: structure, activation, and interactions with fibrinogen and fibrin. Ann N Y Acad Sci 2001; 936: 291–311. 6 Bajzar L, Manuel R, Nesheim ME. Purification and characterization of TAFI, a thrombin- activable fibrinolysis inhibitor. J Biol Chem 1995; 270: 14477–84. 7 Wolberg AS, Campbell RA. Thrombin generation, fibrin clot formation and hemostasis. Transfus Apher Sci 2008; 38: 15–23. 8 Kim PY, Tieu LD, Stafford AR, Fredenburgh JC, Weitz JI. A high affinity interaction of plasminogen with fibrin is not essential for efficient activation by tissue-type plasminogen activator. J Biol Chem 2012; 287: 4652–61. 9 Talbot K, Meixner SC, Pryzdial ELG. Enhanced fibrinolysis by proteolysed coagulation factor Xa. Biochim Biophys Acta 2010; 1804: 723–30. 10 Cesarman-Maus G, Hajjar KA. Molecular mechanisms of fibrinolysis. Br J Haematol 2005; 129: 307–21. 11 Longstaff C, Thelwell C, Williams SC, Silva MMCG, Szabo L, Kolev K. The interplay between tissue plasminogen activator domains and fibrin structures in the regulation of fibrinolysis: kinetic and microscopic studies. Blood 2011; 117: 661–8. 12 Longstaff C, Kolev K. Basic mechanisms and regulation of fibrinolysis. J Thromb Haemost 2015; 13(Suppl. 1): S98–105. 13 Pryzdial ELG, Bajzar L, Nesheim ME. Prothrombinase components can accelerate tissue plasminogen activator-catalyzed plasminogen activation. J Biol Chem 1995; 270: 17871– 7. 14 Pryzdial ELG, Lavigne N, Dupuis N, Kessler G. Plasmin con verts factor X from coagulation zymogen to fibrinolysis cofactor. J Biol Chem 1999; 274: 8500–5. 15 Talbot K, Meixner SC, Pryzdial ELG. Proteolytic modulation of factor Xa–antithrombin complex enhances fibrinolysis in plasma. Biochim Biophys Acta 2013; 1834: 989–95. 16 Grundy JE, Hancock MA, Meixner SC, Mackenzie CR, Koschinsky ML, Pryzdial ELG. Plasminogen binds to plasmin-modulated factor Xa by Ca(2+)- and C-terminal lysine- dependent and -independent interactions. Thromb Haemostasis 2007; 97: 38–44. 17 Grundy JE, Lavigne N, Hirama T, MacKenzie CR, Pryzdial ELG. Binding of plasminogen and tissue plasminogen activator to plasmin-modulated factor X and factor Xa. Biochemistry 2001; 40: 6293–302. 18 Pryzdial ELG, Kessler GE. Autoproteolysis or plasmin-mediated cleavage of factor Xa exposes a plasminogen binding site and inhibits coagulation. J Biol Chem 1996; 271: 16614–20. 19 Semeraro F, Incampo F, Ammollo CT, Dellanoce C, Paoletti O, Testa S, Colucci M. Dabigatran but not rivaroxaban or apixaban treatment decreases fibrinolytic resistance in patients with atrial fibrillation. Thromb Res 2016; 138: 22–9. 20 Ammollo CT, Semeraro F, Incampo F, Semeraro N, Colucci M. Dabigatran enhances clot susceptibility to fibrinolysis by mechanisms dependent on and independent of thrombin-activatable fibrinolysis inhibitor. J Thromb Haemost 2010; 8: 790–8. 21 Varin R, Mirshahi S, Mirshahi P, Klein C, Jamshedov J, Chidiac J, Perzborn E, Mirshahi M, Soria C, Soria J. Whole blood clots are more resistant to lysis than plasma clots – greater efficacy of rivaroxaban. Thromb Res 2013; 131: E100–9. 22 Martin AC, Gouin-Thibault I, Siguret V, Mordohay A, Samama CM, Gaussem P, Le BB, Godier A. Multimodal assessment of non-specific hemostatic agents for apixaban reversal. J Thromb Haemost 2015; 13: 426–36. 23 Escolar G, Fernandez-Gallego V, Arellano-Rodrigo E, Roquer J, Reverter JC, Sanz VV, Molina P, Lopez-Vilchez I, Diaz-Ricart M, Galan AM. Reversal of apixaban induced alterations in hemostasis by different coagulation factor concentrates: significance of studies in vitro with circulating human blood. PLoS ONE 2013; 8: e78696. 24 Tripodi A, Padovan L, Veena C, Scalambrino E, Testa S, Peyvandi F. How the direct oral anticoagulant apixaban affects thrombin generation parameters. Thromb Res 2015; 135: 1186– 90. 25 Frost C, Wang J, Nepal S, Schuster A, Barrett YC, Mosqueda-Garcia R, Reeves RA, LaCreta F. Apixaban, an oral, direct factor Xa inhibitor: single dose safety, pharmacokinetics, pharmacodynamics and food effect in healthy subjects. Br J Clin Pharmacol 2013; 75: 476–87. 26 Dale BJ, Chan NC, Eikelboom JW. Laboratory measurement of the direct oral anticoagulants. Br J Haematol 2016; 172: 315–36. 27 Bajzar L, Nesheim M. The effect of activated protein C on fibrinolysis in cell-free plasma can be attributed specifically to attenuation of prothrombin activation. J Biol Chem 1993; 268: 8608–16. 28 Simpson ML, Goldenberg NA, Jacobson JL, Bombardier CG, Hathaway WE, Manco- Johnson MJ. Simultaneous thrombin and plasmin generation capacities in normal and abnormal states of coagulation and fibrinolysis in children and adults. Thromb Res 2011; 127: 317–23. 29 Matsumoto T, Nogami K, Shima M. Simultaneous measurement of thrombin and plasmin generation to assess the interplay between coagulation and fibrinolysis. Thromb Haemost 2013; 110: 761–8. 30 Broze GJ Jr, Higuchi DA. Coagulation-dependent inhibition of fibrinolysis: role of carboxypeptidase-U and the premature lysis of clots from hemophilic plasma. Blood 1996; 88: 3815–23. 31 Orfeo T, Gissel M, Butenas S, Undas A, Brummel-Ziedins KE, Mann KG. Anticoagulants and the propagation phase of thrombin generation. PLoS ONE 2011; 6: e27852. 32 Weitz JI. New oral anticoagulants: a view from the laboratory. Am J Hematol 2012; 87(Suppl. 1): S133–6. 33 Varin R, Mirshahi S, Mirshahi P, Kierzek G, Sebaoun D, Mishal Z, Vannier JP, Yvonne BJ, Simoneau G, Soria C, Soria J. Clot structure modification by fondaparinux and consequence on fibrinolysis: a new mechanism of antithrombotic activity. Thromb Haemost 2007; 97: 27–31. 34 Schneider M, Nesheim M. Reversible inhibitors of TAFIa both promote and inhibit fibrinolysis. J Thromb Haemost 2002; 1: 147–54. 35 Brummel KE, Paradis SG, Butenas S, Mann KG. Thrombin functions during tissue factor- induced blood coagulation. Blood 2005; 100: 148–52. 36 Yao X, Abraham NS, Sangaralingham LR, Bellolio MF, McBane RD, Shah ND, Noseworthy PA. Effectiveness and safety of dabigatran, rivaroxaban, and apixaban versus warfarin in nonvalvular atrial fibrillation. J Am Heart Assoc 2016; 5: e003725. 37 Patel MR, Mahaffey KW, Garg J, Pan G, Singer DE, Hacke W, Breithardt G, Halperin JL, Hankey GJ, Piccini JP, Becker RC, Nessel CC, Paolini JF, Berkowitz SD, Fox KA, Califf RM. Rivaroxaban versus warfarin in nonvalvular atrial fibrillation. N Engl J Med 2011; 365: 883–91. 38 Granger CB, Alexander JH, McMurray JJ, Lopes RD, Hylek EM, Hanna M, Al-Khalidi HR, Ansell J, Atar D, Avezum A, Bahit MC, Diaz R, Easton JD, Ezekowitz JA, Flaker G, Garcia D, Geraldes M, Gersh BJ, Golitsyn S, Goto S, et al. Apixaban versus warfarin in patients with atrial fibrillation. N Engl J Med 2011; 365: 981–92. 39 Pinto DJP, Orwat MJ, Koch S, Rossi KA, Alexander RS, Smallwood A, Wong PC, Rendina AR, Luettgen JM, Knabb RM, He K, Xin BM, Wexler RR, Lam PYS. Discovery of 1-(4-methoxyphenyl)-7-oxo-6-(4-(2-oxopiperidin-1-yl)phenyl)-4,5,6,7-tetrahydro-1H- pyrazolo[3,4-c]pyridine-3-carboxamide (apixaban, BMS-562247), a highly potent, selective, efficacious, and orally bioavailable inhibitor of blood coagulation factor Xa. J Med Chem 2007; 50: 5339–56. 40 Roehrig S, Straub A, Pohlmann J, Lampe T, Pernerstorfer J, Schlemmer KH, Reinemer P, Perzborn E. Discovery of the novel antithrombotic agent 5-chloro-N-({(5S)-2-oxo-3- [4-(3-oxomorpholin-4-yl)phenyl]-1,3-oxazolidin-5-yl}methyl)thiophene-2- carboxamide (BAY 59-7939): an oral, direct factor Xa inhibitor. J Med Chem 2005; 48: 5900–8. 41 Safouris A, Triantafyllou N, Parissis J, Tsivgoulis G. The case for dosing dabigatran: how tailoring dose to patient renal function, weight and age could improve the benefit-risk ratio. Ther Adv Neurol Disord 2015; 8: 245–54. 42 Iino M, Takeya H, Takemitsu T, Nakagaki T, Gabazza EC, Suzuki K. Characterization of the binding of factor Xa to fibrinogen/fibrin derivatives and localization of the factor Xa binding site on fibrinogen. Eur J Biochem 1995; 232: 90–7. 43 Larsen TB, Skjoth F, Nielsen PB, Kjaeldgaard JN, Lip GY. Comparative effectiveness and safety of non-vitamin K antagonist oral anticoagulants and warfarin in patients with atrial fibrillation: propensity weighted nationwide cohort study. BMJ 2016; 353: i3189. 44 Spronk HMH, de Jong AM, Crijns HJ, Schotten U, Van Gelder IC, ten Cate H. Pleiotropic effects of factor Xa and thrombin: what to expect from novel anticoagulants. Cardiovasc Res 2014; 101: 344–51. 45 Kahn SR, Shrier I, Julian JA, Ducruet T, Arsenault L, Miron M-J, Toussin A, Desmarais S, Joyal F, Kassis J, Solymass S, Desjardins L, Lamping DL, Johri M, Ginsberg JS. Determinants and time course of the postthrombotic syndrome after deep venous thrombosis. Intern Med 2008; 149: 698–707. 46 Kahn SR, Ginsberg JS. Relationship between deep venous thrombosis and the postthrombotic syndrome. Arch Intern Med 2004; 164: 17–26. 47 Crisan S, Vesa S, Pestrea C, Herghea D, Vornicescu D, Chirila M, Crisan IM. Chronic thrombotic scarring in patients with acute deep venous thrombosis of the lower limbs. Med Ultrason 2010; 12: 114–19. 48 Bouman AC, Smits JJ, Ten CH, Ten Cate-Hoek AJ. Markers of coagulation, fibrinolysis and inflammation in relation to post-thrombotic syndrome. J Thromb Haemost 2012; 10: 1532–8. 49 Cheung YW, Middeldorp S, Prins MH, Pap AF, Lensing AW, Ten Cate-Hoek AJ, Villalta S, Milan M, Beyer-Westendorf J, Verhamme P, Bauersachs RM, Prandoni P. Post- thrombotic syndrome in patients treated with rivaroxaban or enoxaparin/vitamin K antagonists for acute deep-vein thrombosis. A post-hoc analysis. Thromb Haemost 2016; 116: 733–8.

Figures

Fig. 1. Cleavage of activated factor X (FXa) by plasmin (Pn) is altered by active site inhibitors and anionic phospholipid-containing membrane (PL) binding. (A) When intact FXa (FXa) is not bound to PL, it is sequentially cleaved by plasmin to FXa and Xa40. Neither bind to plasminogen (Pg), and they have no discernible fibrinolytic function. (B) In contrast, the binding of FXa to PL alters cleavage by plasmin, and the FXa produced interacts with Pg and is fibrinolytic. Plasmin cleaves this form of FXa to Xa33/13, and both associate with Pg. Although Xa33/13 enhances fibrinolysis of purified fibrin, this function is rapidly degraded in plasma. (C) When FXa is inhibited by chloromethyl ketone, the active site-modified FXa (FXai) produced by plasmin is not further cleaved to Xa33/13, and enhances plasma fibrinolysis. Regardless of whether FXa is produced when bound to PL, clotting function is maintained, but Xa40 and Xa33/13 have negligible clotting activity in plasma (data not shown). Gla, -carboxyglutamic acid-containing domain (PL-binding site); I, active site inhibitor; ▲, plasmin cleavage site determined by N-terminal amino acid sequencing; , plasmin cleavage site inferred by N-terminal amino acid sequencing and lack of Pg binding [47]; ◊, inferred from electrophoretic mobility and Pg binding [2,9,13–17]).

Fig. 2. FXa persistence correlates with enhanced fibrinolysis in rivaroxaban-treated patient plasma. (A) Western blots showing the effects of rivaroxaban or apixaban on activated factor X (FXa) fragmentation in normal pooled plasma supplemented with tissue-type plasminogen activator (t-PA) (50 nM) and induced to clot with Innovin. (B) As in (A), except that plasma was obtained from patients taking dabigatran. (C) As in (A), except that plasma was obtained from patients taking rivaroxaban. (B, C) The levels of direct oral anticoagulants in patient plasma are indicated. (D) The rivaroxaban- treated patient or non-treated control pooled plasma samples from (A) and (C) were supplemented with t-PA, and clotting was initiated with excess thrombin (10 nM) to derive the time to half clot lysis as determined by turbidity. Each point represents the mean of three independent experiments conducted in triplicate ± standard error of the mean. (E) Percentage FXa or FXa relative to total FXa in patient plasma at 60 min was measured by densitometry of bands in (C) and plotted against the half clot lysis times shown in (D). r = 0.93, P = 0.003.

Fig. 3. Direct oral anticoagulants (DOACs) enhance normal pooled plasma clot lysis. (A) Thrombin (10 nM)-induced normal pooled plasma clots were supplemented with tissue- type plasminogen activator (t-PA) (30 pM). The extents of clot formation and dissolution were followed by turbidity in clots containing rivaroxaban or apixaban at the maximum plasma concentration, or buffer (t-PA alone). Solid line: n = 3 averaged data ± standard error of the mean (SEM) indicated by gray shading. (B) Under the same conditions as in (A), DOACs were titrated, and the area under the fibrinolysis profile was determined. n = 3 ± SEM (arbitrary fit of data). (C) Thrombin activity was monitored during fibrinolysis under the conditions in (A) by use of a thrombin fluorogenic substrate in the presence or absence of various combinations of added activated factor X (FXa) and DOAC. n = 3 ± SEM. apix, apixaban; riva, rivaroxaban. [Color figure can be viewed at wileyonlinelibrary.com]

Fig. 4. Inhibition of activated thrombin-activatable fibrinolysis inhibitor does not affect enhanced fibrinolysis caused by direct oral anticoagulants. Thrombin-induced plasma fibrinolysis was performed as in Fig. 2 in the absence or presence of potato tuber carboxypeptidase inhibitor (PTCI) (2 M). Each point (●) represents the area under the curve of fibrinolysis assays, and the mean is indicated (∎). n = 4 ± standard error of the mean. t-PA, tissue-type plasminogen activator.

Fig. 5. Fibrin crosslinking is not affected by direct oral anticoagulants (DOACs). Clots were formed by use of the same final concentrations as in Fig. 3, but tissue-type plasminogen activator was omitted. In the presence or absence of DOACs, clots were formed within 1 min, and a band appeared with an apparent molecular weight consistent with crosslinked -chain dimer (–). HMW, high molecular weight.

Fig. 6. Fibrin fiber diameter and porosity are not altered by direct oral anticoagulants. Clots were formed as in Fig. 2, and were then imaged by scanning electron microscopy at three random locations within each clot. (A) Platinum/palladium scanning electron microscope images from one location. (B) The average diameter of non-superimposed fibers was measured. Standard deviation was calculated from the average of 30 measurements (10 fibers measured per location). (C) Porosity was measured from binary images. n = 3 ± standard deviation.

Fig. 7. FX is required for direct oral anticoagulants to enhance plasma clot lysis and plasmin generation. (A) Thrombin (10 nM)-induced clots were formed in FX-deficient plasma (FXDP) purified FX (170 nM) supplemented with 30 pM tissue-type plasminogen activator (t-PA. The extents of clot formation and dissolution were followed by turbidity as in Fig. 2A. n = 3 (average is dark line) standard deviation (SD) (gray shading). (B) The area under the curve was derived from plasma fibrinolysis profiles obtained in the absence () or presence of rivaroxaban (∎) or apixaban (Δ) in FXDP ± FX. Independent experiments were conducted in triplicate, n = 3 ± standard error of the mean. (C) Plasmin activity was monitored as in (A), by use of a plasmin fluorogenic substrate: (a) normal plasma; (b) FXDP; (c) FXDP supplemented with 170 nM FX. n = 3 ± SD. apix, apixaban; riva, rivaroxaban.

Fig. 8. Plasmin generation in the presence of direct oral anticoagulants (DOACs) correlates with altered activated factor X (FXa) fragmentation. Activation of purified plasminogen (Pg) (0.5 M) by tissue-type plasminogen activator (t-PA) (10 nM) was initiated in the presence of anionic phospholipid-containing membrane (PL) (50 M) and calcium (2 mM), with or without FXa (0.1 M), rivaroxaban (0.57 M), or apixaban (0.44 M). (A) Non-reduced anti-human FX western blot and [125I]Pg ligand blots. Markers were positioned by the use of purified FXa fragments and molecular weight markers (data not shown). (B) In the absence of t-PA, purified FXa was treated with plasmin, in the presence of PL and calcium with or without DOACs. Samples were withdrawn at the indicated times, subjected to SDS-PAGE, and stained with Coomassie Blue. (C) Plasmin production was monitored by S2251 chromogenic substrate hydrolysis in samples from experiments conducted as in (A): FXa (); t-PA + FXa + rivaroxaban (▲); t-PA + FXa + apixaban (∎); t-PA + rivaroxaban (Δ); t-PA + apixaban (□); t-PA alone (●). FXa ± DOAC in the absence of t-PA was also evaluated, and found to be at baseline (data not shown). n = 3 ± standard deviation.