ALBERT

Description: Following a vascular injury, factor VIII (FVIII) is rapidly activated by thrombin cleavage at arginine (Arg) 372, 740 and 1689. Activated FVIII, an heterotrimer composed by the association of A1, A2 and A3-C1-C2 domains, is rapidly degraded to limit thrombosis risk. Two main phenomenons that account for the disappearence of FVIIIa consist of an intrinsic dissociation of the trimer due to the loss of A2 domain, and the cleavage of the molecule by activated protein C (APC). APC cleaves FVIIIa at Arg 336 and 562. Mutant FVIII molecules were already generated, with one or two arginines substituted, and a subsequent APC cleavage diminished. Since the thrombotic potential of high plasma levels of FVIII has been described, we aimed to generate new factor VIII molecules where the APC cleavage was only modulated. The ultimate goal being to increase in vivo the FVIIIa half-life. Among the two APC cleavage sites, the sequence around the site 562 was the most conserved between species. This region was therefore chosen to be modified with a prior verification that the amino-acids to be modified were not described in any hemophilic phenotype. Subsequently, the six following mutations were realized in a BDD-FVIII cDNA: Q561N, G563A, N564D, N564A, N564Q and I566M as well as the controls R336I, R562K and R336I+R562K. The constructs were transiently expressed in BHK cells to determine the specific activities of the corresponding molecules. The mutant specific activities, as determined by one- and two-stage clotting assays, ranged from 40 to 94 % of the wild-type except for G563A that was almost inactive. CHO clones expressing FVIII molecules were obtained. The mutants and control FVIII molecules were produced, partially purified on heparin column and further analyzed. The comparative specific activities in a two-stage clotting assay were the following (+/− SD; n=6): BDD-FVIII 100 %, Q561N 105 % +/− 45, G563A 6 % +/− 6, N564D 50 % +/− 23, N564A 45 % +/− 21, N564Q 80 % +/− 26, I566M 100 % +/− 41. The one-stage clotting assay gave identical results than the two-stage assay for each mutant. The mutants were then activated by thrombin (1:1) and the occurrence of FVIII activity was monitored. A 20- to 25-fold increase in FVIII activity was measured within 2 minutes following the addition of thrombin for all mutants. The mutants, except for the inactive G563A, were then analyzed for their resistance to APC by three different assays: an APC resistance kit (Coatest, Chromogenix), an in vitro assay that measured APC sensitivity of FVIIIa and an immunoblot assay that visualized the cleavage efficiency of the A2 fragment. These three assays confirmed the APC resistance of the previously published R336I, R562K and R336I+R562K, as compared with wild-type FVIII. They also revealed that the mutants N564D, N564A and I566M behaved similarly to the wild-type FVIII whereas Q561N and N564Q mutants were partially resistant to APC. The APC resistance ratio were the following 2.3 +/− 0.3 for BDD-FVIII, 2.1 +/− 0.4 for N564Q, 1.7 +/− 0,1 for Q561N, 1.6 +/− 0.3 for R562K and 1.3 +/− 0.3 for R336I+R562K. The N564Q and Q561N mutants exhibited a profile intermediate between wild type FVIII and R562K regarding of the loss of FVIIIa activity that was confirmed on the immunoblot profile. In conclusion we have generated new factor VIII molecules that retained their full procoagulant function while possessing a reduced sensitivity to APC cleavage.

Description: Thrombin generation assay (TGA) was recently evaluated on a living endothelial-derived cell line (Coll et al. J. Thromb. Haemost. 2013; 11, 1916). This innovative assay brought into an hemostasis assay the cellular components of the anticoagulation pathway (APC and TFPI pathways) as well as a activated cell surface. It might help elucidate the relationship between hemostasis and inflammation in a more complex system. In the aim of evaluating the potential of antithrombin (AT) connecting both processes we set-up a similar assay on human vein endothelial cells (HUVEC). We first demonstrated that thrombin generation can be measured in flat-bottom 96-wells in factor IX-or factor VIII-deficient plasma substituted by either 0.1 or 1 U/ml of FIX or FVIII, respectively. Next, HUVEC were grown and expanded in a complete commercial medium (EndoGRO-LS, Millipore) for no more than 6 passages. Wells were then coated with gelatin 1% and cells seeded at 10,000 cells/well. The binding of plasma-derived AT (Aclotine ®, LFB; France dialyzed in cell culture medium) to HUVEC was demonstrated as being dose- (0.5; 1; 2.5 and 5 U/ml) and time- (0-6 hours) dependent. Saturating conditions were found using 2.5 U/ml AT for a 2h incubation. We also showed that the binding was moderately affected in the presence of heparin at concentrations up to 50 U/ml (loss of 19% of the signal) and not at all following an heparanase I+II+III treatment suggesting that another receptor(s) than cellular heparan sulfates being responsible for this interaction. The effect of AT on coagulation was then compared in the presence of cells or not. To do this cells were grown to confluence, washed with non-supplemented medium and incubated in the presence of the TGA mix (plasma containing AT or not, 0.5 pM Tissue Factor, 4 µM Phospholipids). The reaction was initiated by injection of the FluCa kit thrombin substrate (Stago). In the presence of HUVEC, the efficiency of thrombin generation from a control plasma (Unicalibrator, Stago) was decreased with a lag time increased (from 5.67 min to 6.83 min), the peak height diminished from 204.4 nM thrombin to 150.4 nM and the velocity from 55.8 nM/min to 33.4 nM/min. However, the overall amount of thrombin generated was less affected, diminishing from 1515.5 nM to 1482 nM. These data confirms that the presence of the HUVEC anticoagulants pathways can effectively diminish the thrombin generation. Without cells, the presence of 0.5, 1 or 2 U/ml AT dose-dependently decreased the generation of thrombin from the control plasma. The velocity was decreased by 23.2%, 57.6% and 75.5% and the peak height by 33.5%, 61.5% and 78.8%, respectively. When the same experiment was performed in the presence of HUVEC cells, the concentrations of AT similarly decreased the velocity by 34.2%, 54% and 70 % and the peak height by 39%, 59.1% 74.3%, respectively. There was no difference in the TGA parameters if AT was pre-incubated at the surface of the cells for up to 2h prior the TGA or if it was added extemporaneously. These results indicate that the presence of HUVEC did not modulate the in vitro effect of AT during coagulation. The effect of AT on the cell response during this process are in the process of being investigated with a particular focus on the anti-inflammatory properties of AT. Disclosures Catieau: LFB Biotechnologies: Employment. Chtourou:LFB Biotechnologies: Employment. Plantier:LFB Biotechnologies: Employment.

Description: Abstract 4202 Hemophilia B is characterized by a deficiency of coagulation factor IX (FIX) which is synthesized by the liver. CHO cells commonly used for the production of recombinant coagulation factors have a limited capacity for post-translational modifications leading to incomplete γ-carboxylation, decreased phosphorylation and sulfation of the recombinant coagulation proteins. In addition, studies in animal models and data from clinical studies in hemophilic patients, have shown that recombinant FIX has a lower recovery than plasma-derived preparations. In the present study, the human hepatoma cell line HuH-7 was assessed for the production of a recombinant FIX molecule with post-translational modifications similar to those of the plasma-derived FIX. Methods Human hepatoma cell line HuH-7 was transfected with native FIX cDNA directed by CMV promoter, and cellular clones stably expressing FIX were obtained. Benefix®, Mononine® and a protein extract from human liver were used as controls. Cellular expression was analyzed by northern blot, and FIX production was quantified by ELISA. Procoagulant activities were measured by one-stage clotting and thrombin generation (TGT) assays. Secreted and intracellular FIX were visualized by western blot using an antibody directed against FIX. Intracellular trafficking of FIX was further analyzed using inhibitors of proteasome, lysosome and Golgi pathways. Glycosylation profile was determined after successive digestions of lysates by neuraminidase and N-Glycosydase-F. Results Non transfected HuH-7 cells did not express FIX and northern blot did not show any signal corresponding to FIX mRNA. Transiently transfected HuH-7 cells studies revealed the ability to secrete 0.1μg/ml/24h of FIX. The specific activity, measured by a one-stage clotting assay, reached 200% in comparison with Benefix® or rFIX produced by CHO cells. TGT confirmed the high procoagulant activity showing a 2.5 fold increased thrombin generating capacity in comparison with FIX produced by CHO cells, in the same experimental conditions. The secreted molecule appeared as a single band on SDS-PAGE, and presented the same mobility than Benefix® and Mononine®. Following enzymatic digestion, the migration profile suggested that glycosylation and sialylation were similar to Benefix®. Western blot showed that FIX from cell lysate was retrieved as three bands. The upper band, presenting a lower signal than the two others, had the same molecular weight than the secreted form and the native human FIX extracted from a liver biopsy. The two other bands at different intensities migrated slightly faster than the secreted protein. Brefeldin A significantly blocked intracellular traffic of FIX from ER to the Golgi complex where the protein corresponding to the western blot upper band was dramatically accumulated. No intracellular accumulation of the non-secreted forms was observed, and lysosomal pathway was slightly involved in the degradation of both secreted and non-secreted forms of FIX. Conclusion These data demonstrate that HuH-7 cell line is able to produce in serum free medium a biologically active recombinant FIX with a higher specific activity. The protein is correctly synthesized, follows the classical pathway before secretion and therefore is not retained in the cells. N-linked oligosaccharide sialylation is also correctly performed. This preliminary data suggest that HuH7 cells may represent an effective cellular system for rFIX production. Further biochemical analysis of the post-translational modifications and pharmacokinetic properties are required to better characterize the rFIX molecule secreted by HuH7 cells. Disclosures: No relevant conflicts of interest to declare.

Description: Using Massive Mutagenesis technique® we performed a high-throughput alanine substitution of 206 residues between aminoacids 376 to 649 from factor VIII (FVIII) A2 domain. The pattern of activity and the levels of production of FVIII mutants were assessed following transient expression in COS-1 cells. FVIII mutants that kept at least 50% of wild-type activity were then screened in an inhibitor assay against total immunoglobulin G (IgG) fractions from patients with severe hemophilia A who had developed inhibitory antibodies (n=4; range 6–15 BU/mL) or a non immune IgG as control. In this assay, the cell culture supernatants containing FVIII were incubated in a volume of FVIII-depleted plasma for 1h30 in the presence of IgG. The residual activity was then measured in a chronometric assay. No single mutations were able to significantly allow FVIII to escape inhibitors. Four mutations (S409A, L462A, E507A, L629A) having a tendency to resist to inhibitors were selected and recombined two by two leading to a significant but insufficient resistance to anti-FVIII antibodies. The effect of the mutations was additive since a molecule (FVIII-4A2) combining the 4 substitutions significantly resisted to the inhibitory antibodies. Residual activity of FVIII-4A2 ranged from 8% to up to 82% of the initial activity depending on the inhibitor plasma whereas this residual activity never exceeded 30% for control wild-type FVIII. Following production by CHO cells, purified FVIII-4A2 demonstrated a similar pattern of resistance to the four IgG fractions already assayed. FVIII-4A2 was then assayed against 11 additional unrelated inhibitors (range 3–2662 BU/mL) and displayed also a resistance against 10 out of the 11 IgG fractions. The resistance was in all case only partial in relation with the likely presence of anti-C2 and/or anti-A3-C1 inhibitors within the IgG fractions. As detected in a solid-phase assay, the decrease in inhibitory effect was for some of the IgG fractions partly related to a decrease in their binding capacity. As a control experiment, FVIII-4A2 was poorly recognized by the monoclonal antibody GMA012 directed against the A2 domain. In contrast, the binding to ESH4, an anti-C2 monoclonal antibody was not affected. Such combination of mutations opened the perspective for the generation of a recombinant FVIII molecule that can be used as an effective substitutive FVIII therapy in patients with inhibitors.

Description: Specific anti-activated factor X molecules are currently used for the prevention and the treatment of various thromboembolic disorders. However, despite a growing use of these molecules, they are still devoid of a reliable antidote. Rivaroxaban is a specific anticoagulant targeting activated factor X (FXa). Its potential in inhibiting FXa in vitro and in vivo was demonstrated during the characterization of the molecule. However, the use of FXa to revert the effect of Rivaroxaban in plasma was never studied. To do so the measurement of thrombin generation (TG) using the calibrated automatic thrombinoscope was performed. The ability of purified human FXa (Haematologic Technologies at 10, 50, 100, 500 and 1000 ng/ml) to induce TG in a platelet-poor plasma (PPP) without the induction of the coagulation was first evaluated. There was a FXa dose-dependent TG. The TG profile at concentrations up to 50 ng/ml of FXa was similar than the control profile obtained by a PPP activated by tissue-factor (0.5 pM) and phospholipids. Above 50 ng/ml FXa, the lag time decreased and the endogeneous thrombin potential (ETP) increased with the dose. This pattern revealed the thrombogenic potential of FXa and demonstrated that a dose of 50 ng/ml (or ≈1 nM) FXa was the maximum safer dose identified by this assay. A similar experiment was performed following the activation of plasma with 0.5 pM Tissue-Factor (TF) and 4 µM phospholipids (PL) and adding FXa at 31, 62, 125, 250 and 500 ng/ml. The kinetics of TG in the presence of the different amounts of FXa differed less than when coagulation was not induced. The lag times varies from 3 to 1.83 min with the increasing concentrations of FXa and the peak heights from 120 to 212 nM, being the two most affected parameters. Following the addition of 62 ng/ml (or ≈1.25 nM) FXa, the TG was more effective than a control plasma identically stimulated. Rivaroxaban was then spiked in the PPP at the therapeutic dose of 0.35 µg/ml (or 0.8 µM). Following 0.5 pM TF/4 µM PL stimulation, this dosage completely inhibits the TG. Increasing doses of FXa (31, 62, 125, 250 and 500 ng/ml) were then added and dose-dependently restores the TG. All the parameters of the TG profile were affected by the presence of FXa. The normalization was attained at the dose of 250 ng/ml (or 5 nM) FXa. A similar set of experiment was repeated by activating the plasma with cephalin, used as a model to mimic the initiation of the contact phase coagulation. The pattern of TG was different than following FT/PL activation. With cephalin and for all FXa concentrations identical peak aspects (velocity, ETP and peak height) were obtained differing only by their lag times and times-to-peak. Lag times and times to peak were shortened by the addition of FXa from 10.7 to 3.7 min and 13.2 to 6 min respectively. Plasma were then spiked by Rivaroxaban (0.35 µg/ml) and activated by cephalin in the presence of various concentrations of FXa (31, 62, 125, 250 and 500 ng/ml). A dose-dependent TG was demonstrated with the ETP, the peak height and the velocity increasing with the amount of FXa spiked whereas the lag time and time to peak were shortened. Following the induction by cephalin, the presence of FXa systematically shortened the TG when Rivaroxaban was present or not, when compared to the TG from control plasma. This work aimed to establish the antidote potential of the natural substrate of the anti-Xa molecules and limiting the risk in promoting a thrombotic response. The calibrated thrombin generation assay was used to determine the in vitro efficiency of FXa to induce a normal thrombin generation without primary induction or following an induction by TF/PL or cephalin. The doses of FXa required to normalize coagulation in the presence of Rivaroxaban and following induction were identified. These conditions will now be assessed in vivo in Rivaroxaban treated-mice. In addition of establishing the antidote properties of FXa, this data paved the way to compare its capacities, which are optimal to inhibit such inhibitor, to further antidote in development. Disclosures Grenier: LFB BIotechnologies: Employment. Chtourou:LFB Biotechnologies: Employment. Plantier:LFB Biotechnologies: Employment.

Description: Introduction Eptacog beta [Sevenfact®, coagulation factor VIIa (recombinant)-jncw] (HEMA Biologics and LFB) is a human rFVIIa variant indicated for the treatment and control of bleeding events (BEs) in adults and adolescents with hemophilia A or B and inhibitors. In a phase 3 clinical trial (PERSEPT 1, NCT02020369) two initial dose regimens of eptacog beta (75 µg/kg q3h; and 225 µg/kg followed by 75 µg/kg q3h after 9 hours if necessary) demonstrated hemostatic efficacy at 12 hours (82% and 91% respectively) in inhibitor-related mild or moderate bleeding with a very low incidence of rebleeding. Two additional phase 3 trials have been completed and a phase 4 trial is planned. Emicizumab (Hemlibra®, Genentech) is a bispecific antibody indicated for prophylaxis in persons with hemophilia A (with or without inhibitors) to reduce the frequency of bleeding. Emicizumab cannot treat a BE; instead, persons with hemophilia A or B with inhibitors require the use of a bypassing agent (BPA). In clinical trials and post-marketing surveys, thromboembolic events (TEs), including thrombotic microangiopathy have been observed with the concomitant use of aPCC (Feiba®, Shire) (〉100 U/kg/day for 〉1 day) and emicizumab. These TEs have not been observed with emicizumab and eptacog alfa (rFVIIa, NovoSeven® RT, Novo Nordisk) alone. Aims To determine the in vitro procoagulant activity of eptacog beta alone and in combination with emicizumab utilizing a thrombin generation (TG) assay with hemophilia A (HA) and hemophilia A inhibitor (HAI) plasma. The in vitro TG attributed to the combined products will be used to gauge the potential safety and efficacy of eptacog beta for the treatment of BEs in persons with hemophilia A with inhibitors managed with this type of prophylaxis. Methods TG assays were performed according to the method developed by Hemker et al using a Fluoroskan Ascent fluorometer (Thermo Labsystems). TG curves, peak thrombin, endogenous thrombin potential (ETP), lag time and velocity index were determined for severe HA plasma (10 samples) and HAI plasma (4 high titer samples) using both platelet-poor plasma (phospholipid and tissue factor initiator) and platelet-rich plasma (tissue factor initiator). Eptacog beta concentrations (1 µg/mL and 2 µg/mL) were selected to reflect the approximate peak plasma levels observed in the phase 1b trial [NCT01708564; 75 µg/kg dose (0.72 µg/mL) and 225 µg/kg dose (1.9 µg/mL)]; eptacog beta (5 µg/mL) was also examined. Emicizumab (50 µg/mL and 100 µg/mL) concentrations reflect the approximate maximum steady state levels observed when dosed at 1.5 mg/kg (55 µg/mL) and 6 mg/kg (67 µg/mL). Results When no procoagulants were added to HA and HAI plasma, peak thrombin, ETP and velocity index were reduced, and lag time was increased compared to normal pooled plasma (NPP). Addition of emicizumab (50 or 100 µg/mL) and/or eptacog beta (1, 2, or 5 µg/mL) induced a concentration-dependent increase in peak thrombin, ETP, and velocity index. Notably, the effect was lower than that seen with normal plasma: peak thrombin, ETP and velocity index for all combinations and single agents in HAI plasma were significantly lower than those observed in NPP (P