Haemophilia A is a rare X chromosome linked bleeding disorder in which the deficiency of factor VIII (FVIII) results in impaired haemostasis . Currently, the standard of care is to offer primary prophylactic FVIII infusions to patients with severe haemophilia A [<1% FVIII procoagulant activity (FVIII:C)], which can lead to dramatic decreases in bleeding events for these patients . However, FVIII has a half-life of 10–15 h , necessitating frequent infusions. Extensive efforts have been made to prolong the half-life of FVIII to create extended half-life (EHL) FVIII products with the aim of reducing infusion frequency [4,5]. Successful strategies to date include conjugation of FVIII to polyethylene glycol [6,7] and FVIII fusion to the human immunoglobulin Fc . A further challenge in managing patients with haemophilia A is the development of FVIII inhibitors, a treatment-related complication that can interfere with the haemostatic effect of FVIII replacement therapy, and thereby render the patient vulnerable to bleeding. Bypassing agents, which ‘bypass’ the factors blocked by inhibitors, can be utilized for prophylaxis [activated prothrombin complex concentrate (aPCC, FEIBA; Baxalta Inc, a Takeda company, Lexington, Massachusetts, USA)] and to manage bleeding episodes [aPCC, recombinant activated factor VII (rFVIIa; NovoSeven, Novo Nordisk, Bagsvaerd, Denmark)] in patients with haemophilia and inhibitors [9,10].
Investigational, nonfactor molecules that target inhibitors of coagulation, such as antitissue factor pathway inhibitors (concizumab , PF-06741086 , BAY 1093884 , aptamers [14,15] and peptides ) and siRNA directed against antithrombin III  offer the potential for new avenues of treatment in patients with haemophilia A with or without inhibitors. Regulatory authorities in the USA and Europe have approved emicizumab (Hemlibra, ACE910; Roche, Basel, Switzerland), a bispecific activated factor IX (FIXa) and factor X (FX)-directed antibody, for routine prophylaxis to prevent or reduce the frequency of bleeding episodes in adult and paediatric patients, of the ages newborn and older, with haemophilia A with or without FVIII inhibitors . Emicizumab not only aims to mimic some functions of FVIIIa by bridging the FIXa enzyme and its substrate FX [19–21] but also engages/binds FIX and FXa . Emicizumab normalized the activated partial thromboplastin time (aPTT) in a long-term haemophilia A monkey model , and shortened aPTT and increased thrombin generation in plasma with ex-vivo neutralization of endogenous FVIII in a phase I pharmacokinetics and pharmacodynamics trial in healthy male adults . A phase III trial in patients with haemophilia A without inhibitors showed a significant reduction in bleeding events with emicizumab compared with no prophylaxis or previous FVIII prophylaxis , while in a phase III trial in patients with haemophilia A and FVIII inhibitors, once-weekly emicizumab prophylaxis was associated with a lower rate of bleeding events compared with no prophylaxis .
Unlike EHL FVIII, which is regulated by the same mechanisms as physiological FVIII (e.g. activation, inactivation, direct localization to activated membranes), bispecific nonfactor molecules utilize a nonregulated mode of action to improve coagulation . Information is currently lacking regarding the safety and long-term efficacy of these nonfactor approaches. This is particularly important when considering that the thresholds of FVIII activity resulting in definitions of severe (<1%), moderate (1–5%) or mild (5–40%) haemophilia A are expressed as units of FVIII activity (FVIII:C) . Extensive studies have already been performed to compare the different FVIII products under standardized conditions . Currently, FVIII efficacy in the clinical setting is assessed by either one-stage clotting (aPTT) or chromogenic assay. Global haemostatic assays such as thrombin generation assays provide valuable information during product development and are increasingly used in clinical trials and some comprehensive haemophilia treatment centres [29–33]. Each assay type can be designed and performed in numerous ways (i.e. different reagents, buffers), and instrumentation may differ from laboratory to laboratory, resulting in some variation [33–36].
To address the gap in knowledge on the measurability of emicizumab and emerging nonfactor therapies, we studied the in-vitro thrombin generation profile using different coagulation triggers and measurement of the clotting time, with various aPTT reagents comparing standard clotting time evaluation and clot waveform analysis (CWA). We also used several commercially available chromogenic assays to determine the FVIII-equivalent activity of a FIX(a)/FX(a) bispecific antibody molecule, a sequence identical analogue (SIA) to emicizumab.
Materials and methods
The SIA evaluated in the present study is a purified, bispecific anti-FIX(a), anti-FX(a) antibody produced in HEK293 cells and is a SIA of emicizumab (based on the amino acid sequence published for emicizumab ). SIA was purified from cell supernatant via protein A affinity, cation exchange and size exclusion chromatography. HPLC and SDS-PAGE confirmed biochemical purity. A monospecific bivalent anti-FIX(a) antibody having antigen-binding sequences identical to the FIX(a) arm of emicizumab was purified from cell supernatant via protein A affinity and size exclusion chromatography. At the time these studies were conducted, emicizumab was not commercially available, and thus, SIA served as a surrogate reagent and its development was not intended for clinical use.
Normal reference plasma and individual donor patient plasma or plasma pools (three to six donors) from patients with severe haemophilia A (FVIII <1%), obtained by plasmapheresis in blood centres licensed by the US Food and Drug Administration, were purchased as fresh frozen plasma from George King Bio-Medical Inc. (Overland Park, Kansas, USA). ADVATE [recombinant FVIII (rFVIII)] was from Baxalta US, Inc, a Takeda company (Lexington, Massachusetts, USA). Hemlibra (emicizumab, ACE910) was from Genentech (San Francisco, California, USA).
Chromogenic assays to determine factor VIII activity
A panel of the five most commonly used commercial assay kits for the determination of FVIII activity in patient plasma samples was used to measure SIA: Biophen FVIII:C (Hyphen BioMed, Neuville-Sur-Oise, France), Coatest SP4 FVIII and Coamatic Factor VIII (Chromogenix, Instrumentation Laboratory Company, Bedford, Massachusetts, USA), Siemens FVIII:C (Siemens, Behring, Germany) and Technochrom FVIII:C (Technoclone, Vienna, Austria).
Test samples were prepared by spiking SIA at concentrations of 60, 200 and 600 nmol/l (equivalent to 9, 30 and 90 μg/ml), a range that encompasses the clinically relevant concentration of 30 μg/ml , into severe haemophilia A plasma pools. For comparison of monospecific and bispecific antibodies using Technochrom FVIII:C, samples were diluted in kit buffer. Coatest SP4 FVIII and Technochrom FVIII:C were performed in a microtiter plate; all other methods were performed using the BCS-XP analyser (Siemens).
SIA test samples were diluted using kit buffers. The minimum SIA sample dilutions were 1 : 10 Biophen FVIII:C and Technochrom FVIII:C, 1 : 31 Siemens FVIII:C and 1 : 81 Coatest SP4 FVIII and Coamatic Factor VIII. Samples were serially diluted and tested in multiple, independent test units (Technochrom FVIII:C, n = 4; Biophen FVIII:C, n = 3; Siemens FVIII:C, Coatest SP4 FVIII and Coamatic Factor VIII, n = 2). FXa generation was initiated by the kit substrate and quantified by measuring the change in optical density (OD) at 405 nm over time (min). To determine FVIII-equivalent activity of SIA, a calibration curve was constructed by plotting maximum slope of the absorbance versus time curve values (mOD/min) against known FVIII activity values of standard human plasma (Siemens) calibrated against the WHO standard for FVIII.
Activated partial thromboplastin time assays
aPTT assays were performed with five different reagents [Dapttin (Technoclone), Actin FSL (Siemens), Pathromtin SL (Siemens), Triniclot (Tcoag) and aPTT-SP (Instrumentation Laboratory)] in pooled severe haemophilia A plasma, according to the manufacturers’ instructions, and measured on ACL Pro Elite automated coagulation analyser (Instrumentation Laboratory). SIA was diluted to yield plasma concentrations of 2–1200 nmol/l, and 5–2000 mU/ml rFVIII was used as a reference. Fifty microlitres of severe haemophilia A plasma pool and 50 μl aPTT reagent and 50 μl of sample or reference were incubated for 4 min at 37°C. Clotting was initiated by 50 μl of 25 mmol/l CaCl2 solution and recorded for up to 180 s. Each sample was measured in duplicate. Standard clotting time and CWA  were evaluated as previously described and plotted using GraphPad Prism v6. Linear regression using logarithmic scaling determined FVIII-equivalent activity of SIA, using clotting time as a measurement of response.
Interaction of sequence identical analogue with human and bovine factor IXa and factor X
A Biacore T200 instrument (GE Healthcare, Uppsala, Sweden) determined the binding kinetics of SIA with human or bovine FIXa and FX (Hematologic Technologies Inc., Essex Junction, Vermont, USA). SIA was immobilized on a CM5 biosensor chip (GE Healthcare) to a target immobilization level of 5000 response units. Human or bovine FIXa and FX were diluted to 62.5–2000 nmol/l and applied to the chip with a 30 μl/min constant flow rate. Time for association was 5 min and dissociation occurred over 10 min by changing to buffer (HBS-P+ and 5 mmol/l CaCl2, pH 7.4). Regeneration was with 3 mol/l MgCl2 at a 10 μl/min flow rate.
Calibrated automated thrombography assay
Corn trypsin inhibitor and coagulation factors derived from purified human plasma were obtained from Hematologic Technologies Inc. or Enzyme Research Laboratories (South Bend, Indiana, USA). Human thrombin calibrator, tissue factor (TF, PPP-reagent-LOW), MP reagent and FluCa reagent were obtained from Thrombinoscope BV (Maastricht, The Netherlands). Antihuman FVIII goat plasma was from Shire, a Takeda company.
Thrombin generation was evaluated via calibrated automated thrombography (CAT) as previously described . SIA and FVIII were added to fresh frozen plasma from seven individual donors and two pools of severe haemophilia A plasma and were assayed in two independent repeats with duplicate measurements. Prewarmed (37°C) plasma (80 μl) was triggered by either 1 pmol/l TF (10 μl of PPP-reagent-LOW containing recombinant human TF) or 125 pmol/l FXIa. The final assay well volume was adjusted to 120 μl by adding 10 μl sample buffer to each sample.
Thrombin generation (TG) was initiated with 20 μl FluCa reagent containing fluorogenic substrate and HEPES-buffered CaCl2 (100 mmol/l). Fluorescence was measured with Fluoroskan Ascent (Thermo Labsystems, Helsinki, Finland; 390 nm excitation and 460 nm emission filters) at 37°C for 90 min with measurement intervals of 20 s. Peak thrombin or Cmax, lag time, peak time, endogenous thrombin potential (ETP) and velocity index were calculated using the Thrombinoscope software . For each donor plasma sample, an individual FVIII standard curve was analysed and FVIII-equivalent activity of SIA (%) was calculated using a nonlinear four-parameter fit with interpolation based on FVIII titration.
Factor X activation assay
Human FIXa was obtained from Haematologic Technologies Inc. (Essex Junction, Vermont, USA), human FX was from Hyphen BioMed (Neuville-sur-Oise, France) and chromogenic substrate S-2222 was obtained from Chromogenix (Gothenburg, Sweden). Phospholipids were obtained from Avanti Polar Lipids (Alabaster, Alabama, USA). The liposome formulation (PC/PS 60/40) consisting of 60 mole % DOPC (1,2-dioleoyl-sn-glycero-3-phosphocholine; #850375C), 40 mole % POPS (1-palmitoyl-2-oleoyl-sn-glycero-3-[phospho-L-serine], #830034C) was prepared as a 1 mg/ml suspension in 20 mmol/l Tris, 150 mmol/l NaCl, pH7.4 with 5% (w/v) saccharose by extrusion , using two stacked 400 nm polycarbonate filters in a Lipex Biomembranes device (Vancouver, Canada). After lyophilization and reconstitution in distilled water, the liposomes had a mean diameter of 265 nm as determined by dynamic light scattering (Zetasizer Nano ZS, Malvern Instruments, Worcestershire, UK).
The half-maximal effective concentration (EC50) of SIA binding to FIXa was measured in a FX activation assay using a continuous method. Reaction mixtures contained 0.1, 1 and 10 nmol/l FIXa, 140 nmol/l FX and 20 μmol/l phospholipid vesicles with increasing concentrations of either SIA or monospecific bivalent anti-FIX(a) antibody (1, 6, 20, 60, 200, 600 nmol/l) in HNa/BSA (25 mmol/l HEPES, 175 mmol/l NaCl, 0.1% BSA; pH 7.35) containing 5 mmol/l CaCl2. The reaction mix contained 20 μl human FIXa, 20 μl phospholipids, 20 μl SIA and 20 μl of chromogenic FXa substrate S-2222. After 5 min (37°C) incubation, FX-activation was initiated with 20 μl human FX. Cleavage of chromogenic substrate was measured every 15 s for up to 30 min in a BioTec Powerwave 340 microplate reader (405 nm; BioTek Instruments, Highland Park, Vermont, USA). FXa generation was calculated from a standard curve using defined amounts of known FXa active concentrations. Each experiment was run in independent duplicates for monospecific bivalent anti-FIX(a) antibody or independent triplicates for SIA. By fitting the data to one site-specific binding model, the functional apparent EC50 (or apparent KD) and Vmax were obtained.
Establishing in-vitro similarity of sequence identical analogue to emicizumab
Similarity of SIA to emicizumab was evaluated using four methods, as described above: functional binding to FIXa and FX, enzyme kinetic analysis of antibody-improved FIXa-mediated FX activation (Michaelis Menten Kinetics), aPTT assays in FVIII-deficient plasma using Dapttin and aPTT-SP reagents and CAT assay to detect thrombin generation in FVIII-inhibited plasma.
Unless otherwise specified, all data are represented as mean ± standard deviation. Statistical comparisons of samples were performed by unpaired, two-tailed Student's t-tests. Differences were significant when P value was less than 0.05. Analyses were performed in Microsoft Excel 2010 version 14.0 and/or GraphPad Prism version 6.07.
Measurement of sequence identical analogue by chromogenic assay dependent on test kit components
SIA was analysed using five commercial chromogenic assay test kits commonly used for determination of FVIII activity in patient plasma samples. In three of the five methods (Coatest SP4 FVIII, Coamatic Factor VIII and Siemens FVIII:C), SIA did not improve FX activation. Signals were below the limit of quantitation of the FVIII reference range and not differentiable from the buffer blank. These three test kits contained FIXa and FX of bovine origin. Through real-time measurement of interaction kinetics by Biacore, it was observed that SIA exhibited negligible binding to bovine FIXa and did not bind bovine FX, but could bind human FIXa and FX (Fig. 1a,b), providing explanation for the lack of FXa generation activity of SIA in these three kits.
When employing Technochrom FVIII:C, an assay utilizing human FIXa and bovine FX, SIA induced FXa generation, albeit to a limited extent. The FXa generation activity of 600 nmol/l SIA was equivalent to 32 ± 7 to 35 ± 8 mU/ml FVIII (Table 1). Samples diluted more than 20 times were undetectable using this assay. To further investigate the contribution of the FIX(a) binding arm of SIA on FX activation, a monospecific, bivalent anti-FIX(a) antibody consisting of only the FIX(a)-binding sites of SIA was tested alongside SIA. Comparing the FXa generation activity of the anti-FIX(a) antibody to SIA revealed similar FXa generation profiles (Fig. 2a), indicating that SIA stimulated FXa generation largely through the FIX(a) binding arm alone.
Biophen FVIII:C was the only chromogenic assay that utilized both human FIXa and FX reagents. SIA plasma samples analysed with this method revealed a concentration-dependent increase in FXa generation (Fig. 2b). However, the dose–response curves of FVIII reference and SIA test sample were nonparallel, suggesting that SIA comparability to FVIII is dependent on plasma concentration and dilution factor (Fig. 2b). Nonparallelism was primarily observed within the 10–40× dilution range (Table 1), which is the standard dilution range for patient plasma samples, according to kit instructions.
Measurement of sequence identical analogue via activated partial thromboplastin time assay
SIA activity was measured using aPTT assays in severe haemophilia A plasma with five commercially available reagents. Results from a concentration range of 2–1200 nmol/l SIA were compared with those obtained with rFVIII (Fig. 2c,d). With rFVIII as a reference, equivalent activities of SIA were calculated by two methods: standard clotting time evaluation, and CWA (Table 2). Using a standard clotting time evaluation (Fig. 2c), most SIA concentrations showed out-of-range FVIII-equivalent activity of more than 2000 mU/ml. Thus, this method inaccurately predicts activity at therapeutic concentrations of 60 and 600 nmol/l (9 and 90 μg/ml) SIA. With CWA, a concentration-dependent reduction of clotting time was observed for SIA, which differed substantially depending on the aPTT reagent. CWA resulted in lower FVIII-equivalent activity values for SIA with all reagents (Fig. 2d) compared with standard clotting time evaluation. Nevertheless, FVIII-equivalent activity for 60 nmol/l SIA varied from 49 to 1311 mU/ml (Table 2), depending on the type of reagent. Highest activity levels of SIA were obtained with Dapttin and Actin FSL, independent of the data analysis method. Overall, the results indicate that aPTT assays overestimate the SIA procoagulant effect.
Thrombin generation in haemophilia A plasma confirms variability in measurability of sequence identical analogue
To corroborate previous observations that measurability of SIA depends greatly on assay and triggering conditions, we performed CAT with TF and FXIa triggers in two plasma pools and seven individual donor plasma samples from patients with severe haemophilia A (<1% FVIII:C). Thrombin generation was measured in the plasma of 28 (TF-trigger) or 16 (FXIa-trigger) healthy donors to establish a normal reference range for peak thrombin levels (Cmax). For TF-triggered CAT, mean Cmax of the reference was 84 ± 30 nmol/l (range: 29–148 nmol/l). For FXIa-triggered CAT, mean Cmax of the reference was 502 ± 69 nmol/l (range: 414–683 nmol/l). Haemophilia A plasma was then used to measure FVIII-equivalent activity of SIA.
Thrombin generation triggered with 1 pmol/l TF in haemophilia A plasma spiked with rFVIII showed a concentration-dependent increase in mean peak thrombin levels, with 14 ± 12 nmol/l corresponding to 1% FVIII activity, 36 ± 38 nmol/l at 20% and 75 ± 61 nmol/l at 100% activity (Fig. 3a, Table 3). SIA levels of 60 and 600 nmol/l increased mean peak thrombin from 11 ± 9 nmol/l (blank control) to 16 ± 15 and 35 ± 32 nmol/l, respectively, under the same conditions. Interestingly, when triggered with TF, the time to peak thrombin did not shorten for SIA compared with rFVIII, with the highest concentration of SIA reaching peak thrombin at 28.6 ± 7.0 min, corresponding to times obtained with 0–5% rFVIII (Fig. 3b, Table 3). It is worthwhile to note that the median coefficient of variation (CV%) for these analyses was 101%, mainly owing to plasma samples from two donors with high values in the blank control. In addition, one donor plasma sample also exhibited particular sensitivity to rFVIII. Removing these two donor plasma samples from the analyses would result in 31% median CV%. This phenomenon illustrates the variability present in haemophilia A donor plasma.
Thrombin generation of rFVIII triggered with 125 pmol/l FXIa increased peak thrombin to 165 ± 47 nmol/l at 1%, 371 ± 62 nmol/l at 10% and 424 ± 76 nmol/l at 100% FVIII activity, whereas SIA at 60 and 600 nmol/l increased peak thrombin from undetectable at baseline to 210 ± 71 and 322 ± 68 nmol/l, respectively (Fig. 3c,d, Table 3). Comparatively, even the highest concentration of SIA did not reach peak thrombin levels equivalent to 10% FVIII activity under FXIa-triggered conditions.
Individual donor plasma spiked with SIA demonstrated large interindividual variability in peak thrombin measurements. For 600-nmol/l SIA samples, peak thrombin ranged from 12 to 101 nmol/l for TF-triggered samples and 236–426 nmol/l for FXIa-triggered samples (Fig. 4), equating to 15–35 and 4–8% of FVIII-equivalent activity, respectively (Supplementary Table 1, http://links.lww.com/BCF/A69). Other parameters showed similar disparity and tended to be higher under TF-triggered conditions; the velocity index resembled the thrombin levels, varying from 12 to 23% (TF trigger) and 4 to 8% (FXIa trigger) of FVIII-equivalent activity; ETP varied from 7 to 62% (TF) and from 7 to 37% (FXIa) of FVIII-equivalent activity (Supplementary Table 1, http://links.lww.com/BCF/A69). Wider differences were observed with time-to-peak and lag time (Supplementary Table 1, http://links.lww.com/BCF/A69). Time-to-peak values for FXIa-triggered CAT reached more than 100% FVIII-equivalent activity for half of the patient plasma samples at 200 nmol/l SIA, whereas none of the SIA concentrations tested in TF-triggered CAT reached more than 10% FVIII-equivalent activity (Supplementary Table 1, http://links.lww.com/BCF/A69). The values obtained from donor pools were close to the mean of the seven individual donor plasma samples, confirming the individual patient data (data not shown). Thus, we confirmed that FVIII-equivalency in CAT depends on triggering conditions, parameters analysed and the individual responses to SIA of different plasma donors. Furthermore, low TF levels can generate thrombin partly owing to FIXa activation of FX. Triggering thrombin generation with 0.4 pmol/l TF resulted in 10% of FVIII-equivalent activity compared with 12 and 5% at 1 and 5 pmol/l TF, respectively, demonstrating that different concentrations of TF altered the kinetics of thrombin generation, further contributing to variation if test conditions are not standardized.
Factor X activation by sequence identical analogue depends mainly on factor IXa
In order to determine the main driver of FX activation by SIA, 0.1, 1 or 10 nmol/l FIXa was added to increasing amounts of SIA in continuous FX activation kinetic assays (Fig. 5a). FX activation increased directly in proportion with the concentration of FIXa (Fig. 5a). The functional binding values (EC50) of SIA at 0.1, 1 and 10 nmol/l FIXa were 11.0 ± 2.1, 16.0 ± 3.3 and 51.8 ± 4.3 nmol/l, respectively, which are substantially lower values than those reported in previously published molecular binding data . Furthermore, Vmax increased directly proportionally to FIXa concentration. In the presence of 0.1, 1 and 10 nmol/l FIXa with SIA, 0.055 ± 0.002, 0.581 ± 0.026 and 5.261 ± 0.121 nmol/l FXa was generated, respectively (Fig. 5a). In a parallel experiment, when SIA was substituted with the monospecific bivalent anti-FIX(a) antibody, Vmax values were 0.026 ± 0.005, 0.241 ± 0.156 and 1.848 ± 0.21 nmol/l, respectively, demonstrating the FX-generating activity of the monospecific antibody despite the absence of a FX(a)-binding arm (Fig. 5b). This confirms the above observations that the FIX(a) arm of SIA caused a concentration-dependent increase in FX activation.
Sequence identical analogue demonstrated in-vitro comparability with emicizumab
Similarity of SIA to emicizumab was demonstrated using four different comparison methods: First, in an antibody titration assay to determine functional binding affinity using purified coagulation factors FIXa and FX, SIA and emicizumab resulted in similar levels of functional binding affinity, with EC50 values of 6.3 ± 0.4 and 7.8 ± 0.5 nmol/l, and Vmax values of 1.01 ± 0.02 and 1.08 ± 0.02 nmol/l/min, respectively. Second, kinetics of antibody-improved FIXa-mediated FX activation (Michaelis--Menten analysis) were similar with SIA and emicizumab. Km, Vmax, kcat and kcat/Km values were 0.56 ± 0.13 versus 0.62 ± 0.12 nmol/l, 0.0834 ± 0.0022 versus 0.0769 ± 0.0018 nmol/l/min, 1.324 versus 1.221 min–1 and 2.364 versus 1.969 nmol/l/min, respectively. Third, aPTT assays performed with aPTT-SP and Dapttin in FVIII-deficient plasma from patients with haemophilia A produced similar clotting times with SIA and emicizumab (Fig. 6a). Fourth, no differences in thrombin generation were observed with SIA and emicizumab when two different triggers were used (FXIa, TF) in FVIII-deficient plasma (Fig. 6b). These results demonstrate in-vitro similarity between SIA and emicizumab.
Emerging therapies such as emicizumab constitute a promising treatment for patients with haemophilia A with and without inhibitors; however, an effective test for the accurate monitoring of coagulation parameters has not been established. Current laboratory assays are aimed towards monitoring FVIII replacement therapy and are ill suited to determine emicizumab efficacy. The issues that arise are two-fold: how can we monitor the patient in a routine clinical laboratory setting using current methodology, and which in-vitro assay can best predict the in-vivo efficacy of the nonfactor therapies that replace FVIII in patients with haemophilia A? This study aimed to determine the measurability of a SIA of emicizumab to explore potential methods that might form the basis of monitoring tests.
A comprehensive comparative analysis of SIA versus emicizumab (Hemlibra) demonstrated structural and biochemical similarity between the two agents, showing almost identical parameters for functional binding, improvement of bispecific antibody-induced FIXa-mediated FX activation and enhancement of clotting (aPTT) and haemophilia A plasma thrombin generation. These findings indicate that SIA is an excellent surrogate tool for generating data and can inform conclusions on emicizumab.
FVIII replacement therapy is the standard of care for patients with haemophilia A, and the assays that predict and measure its function are well established. Nonfactor therapies currently in development or entering the market – such as bispecific antibodies [19,26], siRNA directed to reduce antithrombin III  and anti-TFPIs [15,16,40] – can prove effective. However, prediction of their function poses a challenge in clinical practice, especially considering the variation in patients’ thrombotic profiles. This report of in-vitro studies shows that FVIII-equivalence of bispecific antibodies cannot be reliably determined using standard FVIII protocols because SIA activity measured by these assays is strongly influenced by assay type, reagent components, analytical conditions and parameters used for calculation. Thus, the lack of relevant assays and product-specific standards makes it difficult to provide a clinically reliable FVIII-equivalent activity measure and requires the development of new methodology to evaluate the efficacy of any emerging nonfactor therapy.
During the study of five commonly used chromogenic assay kits for measurement of SIA, it was determined that SIA does not support FX activation with bovine factors, as confirmed by limited cross reactivity in direct binding studies. Therefore, kits using only bovine factors are unsuitable to measure SIA-induced FX activation. However, although assays utilizing bovine proteins do not detect emicizumab and cannot be used to measure its activity, chromogenic kits containing bovine proteins may be useful for the measurement of FVIII inhibitory antibodies in patients receiving emicizumab prophylaxis. Biophen FVIII:C is the only commercially available chromogenic kit containing both human FIXa and human FX and was the only assay to obtain FXa generation values with SIA. Owing to the specificity of emicizumab for human target proteins, pharmacodynamic measurements of emicizumab in the phase III HAVEN 1 trial were only measured via central laboratories using the Biophen chromogenic assay, in accordance with guidance from the National Hemophilia Association . In our study, results obtained using chromogenic assays found that, compared with FVIII, the concentration response curve of SIA was nonlinear and nonparallel to the FVIII reference curve. Calculation of FVIII-equivalent activity varied with the dilution of SIA test samples, making it difficult to standardize the test for clinical use. In contrast, testing FVIII activity in patient plasma samples was independent of sample dilution and delivered accurate results for a wide range of FVIII levels. Interestingly, SIA induced measurable FXa generation with the Technochrom FVIII:C assay, which provides bovine FX but human FIXa. This implies that the SIA mode of action in part relies on the interaction of the FIX(a)-binding arm of SIA with human FIXa. This mechanism was further confirmed by demonstrating that a monospecific anti-FIX(a) antibody exhibited the same FXa generation profile as SIA and that higher levels of FIXa dramatically increased FX activity. These results corroborate that the FIX(a)-binding arm alone in the context of SIA substantially contributes to FXa generation in addition to the proposed bridging of FIX(a) and FX(a) .
The aPTT assay is a highly sensitive test for measuring FVIII and SIA. SIA plasma concentrations of 60–600 nmol/l (9–90 μg/ml) resulted in FVIII-equivalent activity beyond the FVIII normal range. It appears that aPTT overestimates SIA clotting effects, which in the clinical setting could result in the misinterpretation of a patient's haemostatic status and affect decision-making by physicians in the administration of comedication to a patient who is bleeding. The discrepancy may be due to different mechanisms of rFVIII and bispecific antibodies; unlike rFVIII, bispecific antibodies do not need an activation step, and complexes of FIXa and bispecific antibody activates FX immediately. The absence of bispecific antibody regulation is compounded by the absence of an inactivation mechanism. aPTT assays analysed using CWA [37,42] allowed for better comparison with FVIII than clotting time, but still varied widely with different aPTT reagents. Therefore, a laboratory's test conditions can result in critical differences in FVIII-equivalent activity of bispecific antibodies, which could lead to the improper monitoring of patients.
Global haemostatic assays such as CAT are common tools for drug development and have gained in popularity as diagnostic tools [43,44]. The CAT assay was used in this study to measure SIA activity intrinsically and extrinsically in comparison with FVIII. Spiking 600 nmol/l SIA in haemophilia A plasma yielded values of around less than 1 to more than 100% FVIII-equivalent activity, depending on the trigger, the donor plasma and the parameter used for calculation (peak thrombin, ETP, velocity index, time-to-peak). Individual donor plasma affected CAT parameter results of SIA, indicating high interpatient variability when evaluating SIA. Peak thrombin FVIII-equivalent activity of SIA was approximately three times higher when the assay used TF trigger instead of FXIa trigger. Peak thrombin with SIA when triggered by FXIa, even at the concentration reported as the therapeutic range for emicizumab, only reached 4–8% FVIII-equivalent activity. These values for FVIII-equivalent activity are lower than those reported by Kitazawa et al. (14.6%) , emphasizing the dependency on assay conditions and need for standardization.
Taken together, the SIA activity profiles generated from different methodologies strongly suggest that the root of variability in SIA activity measurements and the difficulty in establishing FVIII-equivalence lies in the varying amounts of FIXa present or generated in the various methods and with different triggers. FVIII concentration in plasma is only approximately 0.7 nmol/l, whereas FIX is overabundant. Thus, FVIIIa is the limiting factor in the formation of FXase with FIXa, and FX activation is controlled by the FVIIIa concentration generated in individual assay systems. In these experiments, SIA was present at a much higher concentration than FIXa, making FIXa the limiting factor in the reaction; this resulted in the dependency of SIA measurements on FIXa levels. This also explains the phenomenon of aPTT overestimating SIA activity compared with FVIII. All generated FIXa is saturated by SIA and boosts the activity of SIA to shorten clotting time. Similarly, as FXIa-triggered CAT generates more FIXa, the greater availability of FIXa substrate resulted in shorter time-to-peak and higher peak thrombin compared with the TF-triggered CAT assay, which has less FIXa. The shift in limiting factors in the assays prevents the establishment of equivalent activity values between FVIII and SIA.
These findings indicate that analysing SIA and converting current standard protocols to determine the FVIII-equivalent activity of SIA is challenging and perhaps not possible owing to its mechanism and unregulated mode of action. Our findings with SIA are in agreement with those from a recent study evaluating the effects and interferences of emicizumab on coagulation assays . Emicizumab exhibited strong interference with several aPTT-based assays that use an intrinsic pathway trigger, including the one-stage FVIII activity assay, and weak interference with PT and PT-derived fibrinogen assays, supporting the recommendation that these assays should not be used in emicizumab-treated patients. Similar to findings from our study, no FVIII activity was observed with emicizumab in the chromogenic assay for FVIII activity utilizing bovine reagents, while a concentration-dependent effect of emicizumab was observed with the assay containing human reagents . However, in our study, the concentration-dependent effect was nonparallel with FVIII reference, leading to difficulty in standardization for clinical use. The present study evaluated an emicizumab SIA, rather than the medical product itself; however, we have demonstrated structural and biochemical similarity between the two agents. Furthermore, our study included a broader range of assays and reagents, over a broad range of concentrations, to evaluate the suitability of assays for routine monitoring. In addition, an emicizumab-based monospecific anti FIX(a) helped in characterizing emicizumab's mode of action further. The data indicate that part of the procoagulant activity of emicizumab is expressed by its FIX(a) arm only, although complex formation and steric positioning of reaction partners, FIXa and FX, in the context of bispecific emicizumab, may contribute to its overall procoagulant activity.
This study confirms that the results of standard assays such as chromogenic and aPTT, and research tools such as thrombin generation show large discrepancies when used to analyse emicizumab. Therefore, better methodology and product-specific standard reagents are paramount for the development of nonfactor therapies and their integration into the clinical setting.
Under the direction of the authors, medical writing support was provided by Lindsay Napier, PhD, CMPP, employee of Excel Medical Affairs (Fairfield, Connecticut, USA) and editorial support was provided by Jocelyn Hybiske, PhD, Medical writing and editorial support was funded by Baxalta US Inc, a Takeda company, Lexington, Massachusetts, USA.
The study was funded by Baxalta Innovations GmbH, a Takeda company, Vienna, Austria.
Rudolf Hartmann contributed to the design, conception and interpretation of the work. Tjerk Feenstra, Sabine Knappe and Gerald Schrenk contributed to the design, execution, analysis and interpretation of the work. Friedrich Scheiflinger and Michael Dockal contributed to the conception, design and interpretation of the work. All authors revised the work critically for important intellectual content and all authors approved the final version.
Conflicts of interest
Rudolf Hartmann, Sabine Knappe, Gerald Schrenk, Friedrich Scheiflinger and Michael Dockal are all employees of Baxalta Innovations GmbH, a Takeda company, and hold stock in the company. Tjerk Feenstra was employed by Baxalta Innovations GmbH at the time the work was conducted for this study.
1. Hoyer LW. Hemophilia A. N Engl J Med
2. Coppola A, Di Capua M, Di Minno MN, Di Palo M, Marrone E, Ierano P, et al. Treatment of hemophilia: a review of current advances and ongoing issues. J Blood Med
3. van Dijk K, van der Bom JG, Lenting PJ, de Groot PG, Mauser-Bunschoten EP, Roosendaal G, et al. Factor VIII half-life and clinical phenotype of severe hemophilia A. Haematologica
4. Tiede A. Half-life extended factor VIII for the treatment of hemophilia A. J Thromb Haemost
2015; 13: (Suppl 1): S176–S179.
5. Powell JS. Longer-acting clotting factor concentrates for hemophilia. J Thromb Haemost
2015; 13: (Suppl 1): S167–S175.
6. Baxalta US Inc. ADYNOVATE, Antihemophilic factor (Recombinant), PEGylated lyophilized powder for solution for intravenous injection prescribing information. 2017. http://www.shirecontent.com/PI/PDFs/ADYNOVATE_USA_ENG.pdf
. [Accessed 21 September 2017].
7. Bayer Healthcare LLC. KOVALTRY [Antihemophilic Factor (Recombinant)] lyophilized powder for solution for intravenous injection prescribing information. 2016. Available at: http://labeling.bayerhealthcare.com/html/products/pi/Kovaltry_PI.pdf
. [Accessed 21 Sep 2017].
8. Bioverativ Therapeutics Inc. ELOCTATE [Antihemophilic Factor (Recombinant), Fc Fusion Protein] lyophilized powder for solution for intravenous injection prescribing information. 2017. https://www.eloctate.com/Pdfs/full-prescribing-information.pdf
. [Accessed 21 Sep 2017].
9. Ewing N, Escuriola-Ettingshausen C, Kreuz W. Prophylaxis with FEIBA in paediatric patients with haemophilia A
and inhibitors. Haemophilia
10. Ju HY, Jang HL, Park YS. The efficacy of bypassing agents in surgery of hemophilia patients with inhibitors. Blood Res
11. Hilden I, Lauritzen B, Sorensen BB, Clausen JT, Jespersgaard C, Krogh BO, et al. Hemostatic effect of a monoclonal antibody mAb 2021 blocking the interaction between FXa and TFPI in a rabbit hemophilia model. Blood
12. Cardinal M, Kantaridis C, Zhu T, Sun P, Pittman DD, Murphy JE, et al. A first-in-human study of the safety, tolerability, pharmacokinetics and pharmacodynamics of PF-06741086, an antitissue factor pathway inhibitor mAb, in healthy volunteers. J Thromb Haemost
13. Gu JM, Zhao XY, Schwarz T, Schuhmacher J, Baumann A, Ho E, et al. Mechanistic modeling of the pharmacodynamic and pharmacokinetic relationship of tissue factor pathway inhibitor-neutralizing antibody (BAY 1093884) in cynomolgus monkeys. AAPS J
14. Waters EK, Genga RM, Schwartz MC, Nelson JA, Schaub RG, Olson KA, et al. Aptamer ARC19499 mediates a procoagulant hemostatic effect by inhibiting tissue factor pathway inhibitor. Blood
15. Waters EK, Genga RM, Thomson HA, Kurz JC, Schaub RG, Scheiflinger F, et al. Aptamer BAX 499 mediates inhibition of tissue factor pathway inhibitor via interaction with multiple domains of the protein. J Thromb Haemost
16. Dockal M, Hartmann R, Fries M, Thomassen MC, Heinzmann A, Ehrlich H, et al. Small peptides blocking inhibition of factor Xa and tissue factor-factor VIIa by tissue factor pathway inhibitor (TFPI). J Biol Chem
17. Sehgal A, Barros S, Ivanciu L, Cooley B, Qin J, Racie T, et al. An RNAi therapeutic targeting antithrombin to rebalance the coagulation system and promote hemostasis in hemophilia. Nat Med
18. Genetech Inc. Hemlibra prescribing information. 2018. https://www.accessdata.fda.gov/drugsatfda_docs/label/2018/761083s002s004lbl.pdf
. [Accessed 10 May 2019].
19. Shima M, Hanabusa H, Taki M, Matsushita T, Sato T, Fukutake K, et al. Factor VIII-mimetic function of humanized bispecific
antibody in hemophilia A. N Engl J Med
20. Nogami K. A bispecific
antibody mimicking factor VIII in hemophilia A therapy. Rinsho Ketsueki
21. Nogami K. Bispecific
antibody mimicking factor VIII. Thromb Res
2016; 141: (Suppl 2): S34–S35.
22. Kitazawa T, Esaki K, Tachibana T, Ishii S, Soeda T, Muto A, et al. Factor VIIIa-mimetic cofactor activity of a bispecific
antibody to factors IX/IXa and X/Xa, emicizumab
, depends on its ability to bridge the antigens. Thromb Haemost
23. Muto A, Yoshihashi K, Takeda M, Kitazawa T, Soeda T, Igawa T, et al. Antifactor IXa/X bispecific
antibody ACE910 prevents joint bleeds in a long-term primate model of acquired hemophilia A. Blood
24. Uchida N, Sambe T, Yoneyama K, Fukazawa N, Kawanishi T, Kobayashi S, et al. A first-in-human phase 1 study of ACE910, a novel factor VIII-mimetic bispecific
antibody, in healthy subjects. Blood
25. Mahlangu J, Oldenburg J, Paz-Priel I, Negrier C, Niggli M, Mancuso ME, et al. Emicizumab
prophylaxis in patients who have hemophilia A without inhibitors. N Engl J Med
26. Oldenburg J, Mahlangu JN, Kim B, Schmitt C, Callaghan MU, Young G, et al. Emicizumab
prophylaxis in hemophilia A with inhibitors. N Engl J Med
27. Lenting PJ, Denis CV, Christophe OD. Emicizumab
, a bispecific
antibody recognizing coagulation factors IX and X: how does it actually compare to factor VIII? Blood
28. White GC 2nd, Rosendaal F, Aledort LM, Lusher JM, Rothschild C, Ingerslev J. Definitions in hemophilia. Recommendation of the scientific subcommittee on factor VIII and factor IX of the scientific and standardization committee of the International Society on Thrombosis and Haemostasis. Thromb Haemost
29. Viuff D, Barrowcliffe T, Saugstrup T, Ezban M, Lillicrap D. International comparative field study of N8 evaluating factor VIII assay performance. Haemophilia
30. Trossaert M, Regnault V, Sigaud M, Boisseau P, Fressinaud E, Lecompte T. Mild hemophilia A with factor VIII assay discrepancy: using thrombin generation
assay to assess the bleeding phenotype. J Thromb Haemost
31. Rodgers S, Duncan E. Chromogenic factor VIII assays for improved diagnosis of hemophilia A. Methods Mol Biol
32. Pickering W, Hansen M, Kjalke M, Ezban M. Factor VIII chromogenic assays can be used for potency labeling and postadministration monitoring of N8-GP. J Thromb Haemost
33. Turecek PL, Romeder-Finger S, Apostol C, Bauer A, Crocker-Buque A, Burger DA, et al. A world-wide survey and field study in clinical haemostasis laboratories to evaluate FVIII:C activity assay variability of ADYNOVATE and OBIZUR in comparison with ADVATE. Haemophilia
34. Kitchen S, Kershaw G, Tiefenbacher S. Recombinant to modified factor VIII and factor IX: chromogenic and one-stage assays issues. Haemophilia
2016; 22: (Suppl 5): 72–77.
35. Kenet G, Stenmo CB, Blemings A, Wegert W, Goudemand J, Krause M, et al. Intra-patient variability of thromboelastographic parameters following in vivo and ex vivo administration of recombinant activated factor VII in haemophilia patients. A multicentre, randomised trial. Thromb Haemost
36. Sommer JM, Buyue Y, Bardan S, Peters RT, Jiang H, Kamphaus GD, et al. Comparative field study: impact of laboratory assay variability on the assessment of recombinant factor IX Fc fusion protein (rFIXFc) activity. Thromb Haemost
37. Shima M, Thachil J, Nair SC, Srivastava A, Scientific and Standardization Committee. Towards standardization of clot waveform analysis and recommendations for its clinical applications. J Thromb Haemost
38. Hemker HC, Giesen P, Al Dieri R, Regnault V, de Smedt E, Wagenvoord R, et al. Calibrated automated thrombin generation
measurement in clotting plasma. Pathophysio Haemost Thromb
39. Pasi KJ, Rangarajan S, Georgiev P, Mant T, Creagh MD, Lissitchkov T, et al. Targeting of antithrombin in hemophilia A or B with RNAi therapy. N Engl J Med
40. Chowdary P, Lethagen S, Friedrich U, Brand B, Hay C, Abdul Karim F, et al. Safety and pharmacokinetics of anti-TFPI antibody (concizumab) in healthy volunteers and patients with hemophilia: a randomized first human dose trial. J Thromb Haemost
41. National Hemophilia Association. MASAC update on the approval and availability of the new treatment. 2018. https://www.hemophilia.org/Newsroom/Medical-News/MASAC-Safety-Information-Update-on-Emicizumab-HEMLIBRA
. [Accessed 10 May 2019].
42. Matsumoto T, Nogami K, Tabuchi Y, Yada K, Ogiwara K, Kurono H, et al. Clot waveform analysis using CS-2000i distinguishes between very low and absent levels of factor VIII activity in patients with severe haemophilia A
43. van Veen JJ, Gatt A, Bowyer AE, Cooper PC, Kitchen S, Makris M. Calibrated automated thrombin generation
and modified thromboelastometry in haemophilia A
. Thromb Res
44. Young G, Sorensen B, Dargaud Y, Negrier C, Brummel-Ziedins K, Key NS. Thrombin generation
and whole blood viscoelastic assays in the management of hemophilia: current state of art and future perspectives. Blood
45. Adamkewicz JI, Chen DC, Paz-Priel I. Effects and interferences of emicizumab
, a humanised bispecific
antibody mimicking activated factor VIII cofactor function, on coagulation assays. Thromb Haemost
46. Peyvandi F, Oldenburg J, Friedman KD. A critical appraisal of one-stage and chromogenic assays of factor VIII activity. J Thromb Haemost