EA Inhibits the Trans-Glutamination Activity of Human FXIIIa. Inhibition of human FXIIIa by EA was evaluated by using a modified bi-substrate, fluorescence-based trans-glutamination assay, as described earlier [18–21]. Dansylcadaverine and N,N-dimethyl-casein were used as two substrates, which upon FXIIIa-dependence conjugation show a marked increase in fluorescence at 490–550 nm (λEX = 360 nm). To measure the potency and efficacy of EA, the dose-dependence of FXIIIa inhibition was evaluated using the logistic Eq. 1. The potency of inhibition refers to the IC50 (x-axis), whereas the efficacy refers to the net change in residual FXIIIa activity (ΔY) (y-axis).
The inhibition profile is shown in Fig. 1. EA inhibited human FXIIIa with an IC50 of 105.9 ± 32.5 µM and efficacy of 65.7 ± 11.1%. Iodoacetamide (IAA), a nonselective inhibitor of thiol-containing enzymes, was used as a positive control. It inhibited human FXIIIa in previous studies under identical assay condition with an IC50 of 2.9 µM (efficacy = ~ 100%). Overall, EA is a moderate inhibitor of the trans-glutamination activity of FXIIIa under in vitro settings.
EA Inhibits FXIIIa-Mediated Fibrin(ogen) Polymerization. The effect of EA on fibrin(ogen) polymerization was further investigated by SDS-PAGE as reported earlier for tridegin [18–20]. A solution containing 13 mg/ml fibrinogen and 2.0 µg/mL FXIIIa in TrisHCl buffer of pH 7.4 containing 10 mM CaCl2 was clotted in the presence and absence of human α-thrombin (2.5 µg/mL). The resulting mixture was either incubated with EA (5, 25, 100, 500, 1000, 3000, and 5000 µM) or buffer. The clots were incubated for 24 hrs at room temperature before the addition of denaturing buffer of 25 mM NaH2PO4, 5.7 M urea, 1.9% (w/v) SDS and 1.9% (w/v) DTT and then incubated overnight at 25°C. Samples were boiled in a water bath for 10 min before centrifugation at 12 000 g at 20°C for 3 min; the supernatants were examined by SDS-PAGE on homogeneous 7.5 % cross-linked gels and stained with Coomassie Brilliant Blue. 100 µM IAA was used as a positive control. The first lane contains the protein markers, whereas the second lane contains the cross-linked fibrin(ogen) formed in the presence of 45 µL (13 mg/mL fibrinogen + 2 µg/mL human FXIIIa) and 50 µL human α-thrombin in pH 7.4 TrisHCl buffer (Fig. 2A). The lane shows the monomers α-, β-, and γ- bands (~ 50–60 kDa) as well as the cross-linked proteins including the lighter γ-γ dimers (~ 117 kDa) as well as the heavier α-α polymers ( > ~ 210 kDa). In one hand, the 100 µM IAA completely inhibited the formation of the cross-linked proteins including the lighter γ-γ dimers as well as the heavier α-α polymers. In the other hand, EA concentration-dependently inhibited the formation of the lighter γ-γ dimers as well as the heavier α-α polymers. Using Eq. 1, EA appears to demonstrate different inhibition behavior towards the γ-γ dimerization as well as the heavier α-α polymerization (Fig. 2B). EA inhibited the former with an IC50 of ~ 1177 µM and efficacy of 95%, and it inhibited the latter with an IC50 of ~ 120 µM and efficacy of 82% (Fig. 2B) suggesting that EA is more potent inhibitor of the α-α polymerization of fibrin(ogen). Although such difference in the inhibition behavior of EA toward FXIIIa-mediated dimerization and polymerization of fibrin(ogen) monomers is under investigation, the results indicate that the inhibition activity of EA toward the catalytic activity of human FXIIIa is physiologically relevant.
EA Inhibits FXIIIa-Mediated formation of Fibrin(ogen) – α 2 -AP. An important physiological function of human plasma FXIIIa is to attach the α2-AP to fibrin polymers in the blood clot, and thus, it renders the blood clot less susceptible to hydrolysis by plasmin. We investigated the effect of EA on FXIIIa-mediated formation of fibrin(ogen) – α2-AP by western blot assay, as reported previously [22]. Figure 3 reveals that EA inhibited the formation of fibrin(ogen) – α2-AP complex at a concentration as low as 100 µM supporting the physiological relevance of the action of EA. In theory, this effect makes the blood clot more susceptible to hydrolysis by the fibrinolytic enzyme plasmin, and thus, potentially contributes to the bleeding observed with intravenously administered EA.
EA Does Not Affect Human Plasma Clotting Times. To evaluate the effect of EA on other clotting factors most of which are serine proteases, we measured its effect on the plasma clotting times; activated partial thromboplastin time (APTT) and prothrombin time (PT) using the BBL Fibrosystem fibrometer (Becton − Dickinson, Sparles, MD), as reported in our previous studies [20, 21]. In one hand, any effect on the APTT (intrinsic coagulation pathway & common coagulation pathway) should reflect an effect on either thrombin, FXa, FIXa, FXIa, or FXIIa. In the other hand, prolongation of the PT (extrinsic coagulation pathway & common coagulation pathway) is indicative of an effect on thrombin, FXa, or FVIIa. In these experiments, we used three positive controls: dabigatran, rivaroxaban, and AntiF11 (Fig. 4A-4D). Figures 4A reveals that EA does not affect APTT or PT indicating the lack of an effect on any serine protease enzyme involved in the intrinsic, extrinsic, or common coagulation pathway at the highest concentration tested of 2500 µM. Figure 4B indicates that dabigatran, which is a direct, active site thrombin inhibitor equally affects the APTT and PT. Figure 4C indicates that rivaroxaban, which is a direct, active site FXa inhibitor equally affects the APTT and PT. Figure 4D reveals that the selective AntiF11 antibody only affects FXIa in the intrinsic coagulation pathway (See also Table 1). Important to mention here that EA also lacked any effect on the thrombin time (results are not shown) further supporting its selectivity of function at the highest tested concentration. In fact, the behavior of inhibited FXIII(a) (in this case by EA) in term of lacking changes in APTT, PT, and TT is similar to what have been reported for patients with FXIII genetic deficiency [14, 15]. Overall, these results indicate that EA is likely to be a selective inhibitor of the plasma transglutaminase FXIIIa over other plasma serine proteases of the coagulation pathways.
Table 1. Effects of EA and other anticoagulants on human plasma
clotting times (APTT and PT)
Molecule
|
Target
|
APTT2x
|
PT2x
|
EA
|
FXIIIa
|
>>2200 µM
|
>>2200 µM
|
Dabigatran
|
FIIa
|
0.32 µM
|
0.33 µM
|
Rivaroxaban
|
FXa
|
0.19 µM
|
0.15 µM
|
AntiF11
|
FXIa
|
2.2 µg/mL
|
>>3.5 µg/mL
|
APTT2x: The concentration required to double the clotting times of normal human plasma under APTT conditions. PT2x: The concentration required to double the clotting time of normal human plasma under PT conditions. |
EA Recognizes Cys314 in the Active Site of FXIIIa. To identify a plausible binding mode for EA, we performed molecular docking studies by considering the active site of FXIII(a). The rationale for considering this site is that it contains the catalytic triad of which Cys314 appears to be the most critical residue [16, 17]. The covalent docking of EA was achieved through Michael addition reaction by properly defining the interacting groups for the ligand and the enzyme. Covalent docking feature first docks each ligand to the enzyme to generate a non-covalent pose using Glide (Schrodinger Suite) [23, 24]. Then, the covalent bond is formed between the ligand i.e. EA and the enzyme i.e. the thiol of Cys314 of FXIIIa, and the resulting ligand pose is refined and scored by performing MM-GBSA calculations with the OPLS/AA force field and GB/SA continuum model. The best-docked structure based on the docking score was selected for further analysis of the binding features of EA to FXIIIa. Overall. the molecular modeling studies revealed that the catalytic domain of FXIIIa is a potential binding site for EA with a covalent bond forming between the α,β-unsaturated ketone of EA and the catalytic Cys324 of human FXIIIa. Other potential important interactions are H-bonds between the ketone group of EA and the NH-groups of the side chains of Gln313 and Trp279. The carboxylic acid also potentially establishes H-bond with the side chain of the Asn371 residue. The 2-Cl substituent also forms a halogen-H interaction with the side chain of the Asn371 residue. Although these results are to be experimentally confirmed via crystallography studies and/or mutagenesis studies, however, they further support the concept of FXIIIa inhibition as a potential contributor to the bleeding observed with intravenously administered EA.