NXT007 enhances in vitro coagulation potential in the coexistence of emicizumab in hemophilia A through distinct complex formation

Abstract

Background

NXT007 is a factor (F)VIIIa-mimetic bispecific antibody (BsAb) for hemophilia A (HA), engineered from emicizumab. It is designed to achieve non-hemophiliac levels of plasma coagulation potential and is currently under clinical development. In the future, emicizumab users may switch to NXT007.

Objectives

To investigate the effects of NXT007 in the coexistence of emicizumab on coagulation potential in vitro.

Methods

Coagulation potentials of HA plasma were determined by thrombin generation (TG) assays. NXT007 binding was analyzed by biolayer interferometry (BLI) and immunoblotting. Plasma levels of FIX-BsAb-FX ternary complexes were simulated using dissociation constant values for calculating theoretical coagulation potentials.

Results

In coexistence of emicizumab at 50 μg/mL (therapeutic concentration), NXT007 (0.1-100 μg/mL) dose-dependently increased TG capacity of HA plasma, similar to its effect without emicizumab. At 10-100 μg/mL of NXT007, TG capacity reached levels comparable to NXT007-only conditions. BLI analysis revealed that emicizumab and NXT007 could not simultaneously bind to the same antigens, indicating that only ternary complexes or lower stoichiometry can be formed from FIX(a), FX(a), emicizumab, and NXT007. Immunoblotting analyses confirmed both BsAbs recognized the same antigen domains. Therefore, plasma coagulation potential should be the sum of the contributions of both BsAbs. Our calculations in this manner were concordant with the TG results.

Conclusion

NXT007 enhanced TG capacity of HA plasma in vitro in the coexistence of emicizumab without an abnormal synergistic increase or reduction in coagulation potential, consistent with their competitive antigen binding. These findings suggest that switching from emicizumab to NXT007 can be achieved without emicizumab washout.

Essentials

• •NXT007 is an emicizumab-based factor VIIIa mimetic designed to achieve nonhemophiliac levels of coagulation potential.
• •NXT007 enhances in vitro thrombin generation in the presence of emicizumab in hemophilia A plasma.
• •Both antibodies cannot simultaneously bind to the same antigens and therefore form distinct complexes.
• •Switching from emicizumab to NXT007 could be achieved without requiring emicizumab washout.

1. Introduction

Congenital hemophilia A (HA) is an inherited bleeding disorder caused by a deficiency or dysfunction of coagulation factor (F)VIII [1]. FVIII-replacement therapy has been a long-standing therapeutic intervention, initially on demand and subsequently as prophylaxis using frequent intravenous replacement with FVIII concentrate. Emicizumab, the first FVIII mimetic, was developed and launched in 2017 as an alternative treatment option for HA, overcoming many of the limitations of conventional FVIII-replacement therapy [2,3]. Emicizumab is a humanized IgG₄ bispecific antibody (BsAb) with a common light chain that recognizes the first epidermal growth factor (EGF)-like domain (EGF1 domain) of activated FIX (FIXa) and EGF2 domain of FX. It is designed to mimic the function of FVIIIa by promoting FIXa-catalyzed FX activation on activated platelet surfaces by placing FIXa and FX into spatially appropriate positions. Due to the lack of structural homology with FVIII, emicizumab is effective in persons with HA who have alloantibodies (inhibitors) against FVIII. In clinical studies, emicizumab prophylaxis dramatically decreased bleeding episodes [[4], [5], [6], [7], [8]]. Its high subcutaneous bioavailability and long half-life relative to FVIII products allow for less frequent subcutaneous injections (once weekly to once every 4 weeks) and minimize fluctuations between peak and trough levels, thereby substantially improving the quality of life for persons with HA [9].

NXT007 is a next-generation therapeutic antibody created using the amino acid sequence of emicizumab as a template. It is designed to achieve further improvements in activity and pharmacokinetics to reach non-hemophiliac levels of cofactor activity with a user-friendly dosing regimen [10,11]. NXT007 consists of emicizumab-based engineered heavy chains and newly developed non-common light chains. This adoption of 2 different light chains has expanded the flexibility of amino acid substitutions during the engineering process. Targeted amino acid mutations in the variable regions of both heavy and light chains were made to identify effective mutations that enhance cofactor activity. Additionally, in the Fc region, several amino acid mutations were introduced to enhance binding to the neonatal Fc receptor at acidic pH to improve pharmacokinetics. In nonclinical studies, NXT007 demonstrated in vivo hemostatic activity at 30-fold lower plasma concentrations than emicizumab against ongoing bleeds in nonhuman primates with acquired HA, which was concordant with NXT007’s higher activity to enhance the thrombin generation (TG) capacity of HA plasma in vitro. Combined with pharmacokinetic data from nonhuman primate studies, NXT007 would be expected to maintain non-hemophiliac levels of coagulation potential with monthly subcutaneous dosing. Clinical trials evaluating NXT007’s safety and efficacy in persons with HA are currently underway to clarify its clinical potential (jRCT2080224835, NCT05987449) [12,13].

Since over 20,000 persons with HA have been treated with emicizumab [9], persons with HA may have the opportunity to switch from emicizumab to NXT007 in the future. In such situations, emicizumab and NXT007 may transiently coexist in the blood circulation of persons with HA during treatment transitions. If the combination of these agents exerts any unforeseen influence on coagulation potential, such combination effects cannot be ignored since both agents have much longer half-lives than conventional FVIII products. Understanding the effects of drug coexistence in the blood circulation of persons with HA is important for estimating changes in coagulation potential as well as safety risks to maximize clinical benefits. Since NXT007 was created using emicizumab as a template, we hypothesized that NXT007 would compete with emicizumab in antigen binding and form distinct active complexes, such that abnormal enhancement or reduction of coagulation would not occur when they coexist.

In the present study, we investigated the effect of NXT007 in the coexistence of emicizumab. First, we examined the cofactor activity in the coexistence in vitro using TG assays. Then, to interpret the TG results theoretically, we clarified the antigen binding properties in the coexistence of emicizumab and NXT007 using competitive binding analysis and conducted simulations of FIX-BsAb-FX ternary complex formation in plasma of persons with HA. We believe these data can provide insights into the plasma coagulation potential reflected by TG capacity during the transition from emicizumab to NXT007, helping to estimate the possible efficacy and safety profile.

2. Methods

2.1. Antibodies

NXT007 was produced recombinantly in a Chinese hamster ovary cell line at Chugai Pharmaceutical Co, Ltd (Chugai) [10]. Emicizumab is a marketed therapeutic antibody (Chugai) and is also Chinese hamster ovary cell-derived. Anti-FIX(a) and anti-FX monospecific two-armed IgG₄ antibodies, each having one of the antigen-binding fragments of NXT007 (NXT007 Fab), for immunoblot detection were transiently expressed in HEK293 cells and purified. Anti-FIXa catalytic domain (XB12) and anti-FX activation peptide (SB04) antibodies for biolayer interferometry (BLI) analysis [14] were also HEK293 cell-derived (Supplementary Methods). BsAbs having one anti-FIX(a) or anti-FX monospecific arm and one anti-keyhole limpet hemocyanin (KLH; a protein of nonmammalian origin) arm were also recombinantly prepared to be used for anti-FIX(a) or anti-FX monospecific one-armed antibody equivalents for surface plasmon resonance (SPR) analysis and an enzymatic assay.

2.2. TG assay

TG in HA plasma (FVIII-deficient plasma; George King Bio-Medical) was measured by calibrated automated thrombography (Thrombinoscope BV) using a 96-well plate fluorometer (Thermo Fisher Scientific) as described previously [10]. Briefly, each well in a 96-well plate was filled with 80 μL HA plasma containing emicizumab and/or NXT007, to which 20 μL of low tissue factor (TF)-triggering solution using PPP-reagent LOW (Thrombinoscope BV) or FXIa-triggering solution consisting of 31 pM plasma-derived (pd)FXIa (Enzyme Research Laboratories) and 20 μM synthetic phospholipid was added. For calibration, 20 μL of Thrombin Calibrator (Thrombinoscope BV) was added instead of the triggering solution. To initiate the reaction, 20 μL of FluCa reagent prepared from FluCa kit (Thrombinoscope BV) was dispensed by the instrument as programmed. We analyzed the thrombograms using the instrument’s software. Data were collected in triplicate.

2.3. BLI analysis

Epitope competition between antibodies was analyzed using BLI [14] on an Octet RED384 instrument (Sartorius AG) equipped with 384-well tilted-bottom microplates (Sartorius AG) and streptavidin sensor tips (Sartorius AG). All BLI measurements were performed at 30 °C in Tris-buffered saline (TBS; 50 mM Tris-HCl, 150 mM NaCl, pH 7.4). pdFIXa and pdFX (Enzyme Research Laboratories) were biotinylated using No-Weigh Premeasured Sulfo-NHS-SS-Biotin microtubes (Pierce Biotechnology) according to the manufacturer’s instructions. After buffer exchange and adjustment of the concentration to 1 μg/mL in TBS, the biotinylated antigens were used for immobilization. The degree of biotinylation was determined using the EZ Biotin Quantitation Kit (Pierce Biotechnology). For the competition assay, streptavidin sensor tips were first incubated with TBS for baseline stabilization (60 seconds) and with biotinylated antigen (bio-FIXa or bio-FX, 1 μg/mL) for 120 seconds for immobilization. After baseline stabilization (30 seconds), the primary antibody (emicizumab, NXT007; each at 100 μg/mL, or TBS) was associated for 300 seconds, immediately followed by association of the secondary antibody (NXT007, emicizumab, or control antibodies: XB12 for FIXa and SB04 for FX control; each at 100 μg/mL) for an additional 300 seconds. In the secondary antibody association step, antibody mixtures with primary antibody were applied so that each antibody had a final concentration of 100 μg/mL. Sensorgrams were recorded, and the increase in signal after the addition of the secondary antibody was analyzed. Competition was defined as a negligible increase in binding signal upon the addition of the secondary antibody, indicating that the antibodies recognize overlapping or closely situated epitopes. In contrast, an additive increase in signal indicated that the antibodies recognize distinct, non-overlapping epitopes. To confirm the specificity of the observed competition, control assays were performed using antibodies binding to different domains (XB12 for FIXa, SB04 for FX).

2.4. Coomassie brilliant blue (CBB) staining and immunoblotting analysis

Recombinant Fc fusion forms of the first and second EGF-like domains (EGF1,2), the first EGF-like domain (EGF1), and the second EGF-like domain (EGF2) of FIX and FX (rFIX_EGF1,2–Fc; rFIX_EGF1–Fc; rFIX_EGF2–Fc; rFX_EGF1,2–Fc; rFX_EGF1–Fc; and rFX_EGF2–Fc) were prepared as previously reported [15]. For CBB staining or immunoblotting, we applied 20 pmol or 5 pmol of the above FIX- or FX-related recombinant proteins, respectively, and 40 pmol or 10 pmol of pdFIXa and pdFX (Enzyme Research Laboratories), respectively, to gradient acrylamide gels (4%-20%) and performed sodium dodecyl sulfate polyacrylamide gel electrophoresis under nonreducing conditions. For immunoblotting analyses, we used the anti-FIX(a) or anti-FX monospecific two-armed IgG₄ antibodies having either of the NXT007 Fabs as the test antibody and detected the antigen-antibody binding by horseradish peroxidase-labeled anti-kappa [11] light chain antibody (Abcam) for NXT007 anti-FIXa Fab detection or horseradish peroxidase-labeled anti-lambda [11] light chain antibody (Abcam) for NXT007 anti-FX Fab detection and a peroxidase immunoblotting substrate (Bio-Rad).

2.5. SPR analysis

SPR analysis was performed at 37 °C using a Biacore T200 system (Cytiva). Each antibody was captured onto Sure Protein A (MabSelect SuRe) immobilized on a CM4 sensor chip (Cytiva), followed by injection of recombinant (r)FIX (Benefix; Pfizer), pdFIXa, pdFX, or pdFXa (Enzyme Research Laboratories) over the flow cells as described previously [10]. Kinetic analysis was performed by global fitting of binding data with a 1:1 binding model using Biacore Evaluation Software (Cytiva).

2.6. FXa generation enzymatic assay

The conversion rate of FX to FXa was determined in an enzymatic assay using purified coagulation factors, as described previously [2]. The assay system consisted of 1 nM pdFIXa, 140 nM pdFX, 4 μM synthetic phosphatidylserine-containing phospholipid mixture (10% phosphatidylserine, 60% phosphatidylcholine, and 30% phosphatidylethanolamine) [16], and BsAbs. The FXa generation reaction was carried out at room temperature for 1 minute in TBS containing 1 mM CaCl₂ and 0.1% (wt/vol) bovine serum albumin (pH 7.6) and stopped by adding EDTA. Then, S-2222 chromogenic substrate (Werfen) for FXa was added, and absorbance at 405 nm was measured at intervals to determine the rate of FXa generation. Data were collected in triplicate.

2.7. Dissociation constant-based simulation of FIX-BsAb-FX ternary complex formation

FIX-BsAb-FX ternary complex formation in plasma was simulated based on the concentrations of FIX and FX and the dissociation constant (K_D) using the following equations. C_FIX, C_FX, C_Emi, and C_NXT represent total FIX concentration, total FX concentration, total emicizumab concentration, and total NXT007 concentration. C_FIX was calculated as the sum of free FIX concentration (C_FIXfree), FIX-emicizumab binary complex concentration (C_{FIX_Emi}), FIX-emicizumab-FX ternary complex concentration (C_FIX-Emi-FX), FIX-NXT007 binary complex concentration (C_{FIX_NXT}), and FIX-NXT007-FX ternary complex concentration (C_FIX-NXT-FX). C_FX was calculated in same manner. In the simulations, C_FIX and C_FX were assumed to be 90 nM and 135 nM, which were same settings used in a previous report [15]. C_Emi was defined as the sum of free emicizumab concentration (C_Emifree), FIX-emicizumab binary complex concentration (C_{FIX_Emi}), emicizumab-FX binary complex concentration (C_{Emi_FX}), and FIX-emicizumab-FX ternary complex concentration (C_FIX-Emi-FX). C_NXT was calculated in the same manner.

$${\mathit{C}}_{\mathit{F}\mathit{I}\mathit{X}}={\mathit{C}}_{\mathit{F}\mathit{I}\mathit{X}\mathit{f}\mathit{r}\mathit{e}\mathit{e}}+{\mathit{C}}_{\mathit{F}\mathit{I}\mathit{X}\_\mathit{E}\mathit{m}\mathit{i}}+{\mathit{C}}_{\mathit{F}\mathit{I}\mathit{X}\_\mathit{E}\mathit{m}\mathit{i}\_\mathit{F}\mathit{X}}+{\mathit{C}}_{\mathit{F}\mathit{I}\mathit{X}\_\mathit{N}\mathit{X}\mathit{T}}+{\mathit{C}}_{\mathit{F}\mathit{I}\mathit{X}\_\mathit{N}\mathit{X}\mathit{T}\_\mathit{F}\mathit{X}}$$

$${\mathit{C}}_{\mathit{F}\mathit{X}}={\mathit{C}}_{\mathit{F}\mathit{X}\mathit{f}\mathit{r}\mathit{e}\mathit{e}}+{\mathit{C}}_{\mathit{E}\mathit{m}\mathit{i}\_\mathit{F}\mathit{X}}+{\mathit{C}}_{\mathit{F}\mathit{I}\mathit{X}\_\mathit{E}\mathit{m}\mathit{i}\_\mathit{F}\mathit{X}}+{\mathit{C}}_{\mathit{N}\mathit{X}\mathit{T}\_\mathit{F}\mathit{X}}+{\mathit{C}}_{\mathit{F}\mathit{I}\mathit{X}\_\mathit{N}\mathit{X}\mathit{T}\_\mathit{F}\mathit{X}}$$

$${\mathit{C}}_{\mathit{E}\mathit{m}\mathit{i}}={\mathit{C}}_{\mathit{E}\mathit{m}\mathit{i}\mathit{f}\mathit{r}\mathit{e}\mathit{e}}+{\mathit{C}}_{\mathit{F}\mathit{I}\mathit{X}\_\mathit{E}\mathit{m}\mathit{i}}+{\mathit{C}}_{\mathit{E}\mathit{m}\mathit{i}\_\mathit{F}\mathit{X}}+{\mathit{C}}_{\mathit{F}\mathit{I}\mathit{X}\_\mathit{E}\mathit{m}\mathit{i}\_\mathit{F}\mathit{X}}$$

$${\mathit{C}}_{\mathit{N}\mathit{X}\mathit{T}}={\mathit{C}}_{\mathit{N}\mathit{X}\mathit{T}\mathit{f}\mathit{r}\mathit{e}\mathit{e}}+{\mathit{C}}_{\mathit{F}\mathit{I}\mathit{X}\_\mathit{N}\mathit{X}\mathit{T}}+{\mathit{C}}_{\mathit{N}\mathit{X}\mathit{T}\mathbf{\_}\mathit{F}\mathit{X}}+{\mathit{C}}_{\mathit{F}\mathit{I}\mathit{X}\_\mathit{N}\mathit{X}\mathit{T}\_\mathit{F}\mathit{X}}$$

K_D,Emi,FIX, K_D,Emi,FX, K_D,NXT,FIX, and K_D,NXT,FX represent emicizumab’s K_D value to FIX, emicizumab’s K_D value to FX, NXT007’s K_D value to FIX, and NXT007’s K_D value to FX. All K_D values were taken from previous reports [10,15]. K_D was defined as [target free antibody concentration] × [antibody free target concentration] / [antibody-target complex concentration] in the following equations. It was assumed that the binding of either arm is independent.

$${\mathit{K}}_{\mathit{D},\mathit{E}\mathit{m}\mathit{i},\mathit{F}\mathit{I}\mathit{X}}=\frac{\left({\mathit{C}}_{\mathit{E}\mathit{m}\mathit{i}\mathit{f}\mathit{r}\mathit{e}\mathit{e}}+{\mathit{C}}_{\mathit{E}\mathit{m}\mathit{i}\_\mathit{F}\mathit{X}}\right)\ast {\mathit{C}}_{\mathit{F}\mathit{I}\mathit{X}\mathit{f}\mathit{r}\mathit{e}\mathit{e}}}{{\mathit{C}}_{\mathit{F}\mathit{I}\mathit{X}\_\mathit{E}\mathit{m}\mathit{i}}+{\mathit{C}}_{\mathit{F}\mathit{I}\mathit{X}\_\mathit{E}\mathit{m}\mathit{i}\_\mathit{F}\mathit{X}}}$$

$${\mathit{K}}_{\mathit{D},\mathit{E}\mathit{m}\mathit{i},\mathit{F}\mathit{X}}=\frac{\left({\mathit{C}}_{\mathit{E}\mathit{m}\mathit{i}\mathit{f}\mathit{r}\mathit{e}\mathit{e}}+{\mathit{C}}_{\mathit{E}\mathit{m}\mathit{i}\_\mathit{F}\mathit{I}\mathit{X}}\right)\ast {\mathit{C}}_{\mathit{F}\mathit{X}\mathit{f}\mathit{r}\mathit{e}\mathit{e}}}{{\mathit{C}}_{\mathit{F}\mathit{X}\_\mathit{E}\mathit{m}\mathit{i}}+{\mathit{C}}_{\mathit{F}\mathit{I}\mathit{X}\_\mathit{E}\mathit{m}\mathit{i}\_\mathit{F}\mathit{X}}}$$

$${\mathit{K}}_{\mathit{D},\mathit{N}\mathit{X}\mathit{T},\mathit{F}\mathit{I}\mathit{X}}=\frac{\left({\mathit{C}}_{\mathit{N}\mathit{X}\mathit{T}\mathit{f}\mathit{r}\mathit{e}\mathit{e}}+{\mathit{C}}_{\mathit{N}\mathit{X}\mathit{T}\_\mathit{F}\mathit{X}}\right)\ast {\mathit{C}}_{\mathit{F}\mathit{I}\mathit{X}\mathit{f}\mathit{r}\mathit{e}\mathit{e}}}{{\mathit{C}}_{\mathit{F}\mathit{I}\mathit{X}\_\mathit{N}\mathit{X}\mathit{T}}+{\mathit{C}}_{\mathit{F}\mathit{I}\mathit{X}\_\mathit{N}\mathit{X}\mathit{T}\_\mathit{F}\mathit{X}}}$$

$${\mathit{K}}_{\mathit{D},\mathit{N}\mathit{X}\mathit{T},\mathit{F}\mathit{X}}=\frac{\left({\mathit{C}}_{\mathit{N}\mathit{X}\mathit{T}\mathit{f}\mathit{r}\mathit{e}\mathit{e}}+{\mathit{C}}_{\mathit{N}\mathit{X}\mathit{T}\_\mathit{F}\mathit{I}\mathit{X}}\right)\ast {\mathit{C}}_{\mathit{F}\mathit{X}\mathit{f}\mathit{r}\mathit{e}\mathit{e}}}{{\mathit{C}}_{\mathit{F}\mathit{X}\_\mathit{N}\mathit{X}\mathit{T}}+{\mathit{C}}_{\mathit{F}\mathit{I}\mathit{X}\_\mathit{N}\mathit{X}\mathit{T}\_\mathit{F}\mathit{X}}}$$

2.8. Statistical analysis

The mean ± SD of 3 sets of measurements is presented for the FXa generation activity and TG peak thrombin values.

3. Results

3.1. In vitro effect of NXT007 on TG capacity of HA plasma in the coexistence of emicizumab

Previously, we showed that NXT007 at concentrations >4.4 μg/mL enhanced TG capacity of HA plasma to non-hemophiliac levels compared with FVIII in the TF-triggered in vitro TG assay [10]. This in vitro data also indicated a 30-fold dose reduction from emicizumab to NXT007, which was consistently observed in the in vivo study using nonhuman primates with acquired hemophilia A [10]. Therefore, we assumed that the in vitro TF-triggered TG data of emicizumab and NXT007 may predict their in vivo activity. In this study, we assessed changes of in vitro TF-triggered TG capacity of HA plasma when spiking NXT007 in the coexistence of emicizumab at the clinical therapeutic plasma concentration (50 μg/mL) and at a lower concentration (10 μg/mL). As shown in Figure 1A, the addition of NXT007 at concentrations of 0.1 to 100 μg/mL dose-dependently enhanced peak height values of HA plasma containing emicizumab at 10 and 50 μg/mL. In either case, NXT007’s dose dependence for enhancing peak height in the coexistence of emicizumab was similar to that observed in the absence of emicizumab. At NXT007 concentrations of 10 and 100 μg/mL, absolute peak height values were not meaningfully altered by the coexistence of emicizumab at 10 or 50 μg/mL. In addition, even in the coexistence of emicizumab at 300 μg/mL, which far exceeds the therapeutic plasma concentration, NXT007’s dose-dependent enhancement of peak height was maintained, although the value was slightly lower than the NXT007-only condition with 100 μg/mL of NXT007. To ensure the consistency of the dose dependence of NXT007’s effects, we also examined it using a FXIa-triggered TG assay and observed a similar dose dependence (Figure 1B). These TG assay results suggest that dose-dependent enhancement of TG capacity of HA plasma by NXT007 can be observed even in the coexistence of emicizumab at the clinical therapeutic plasma concentration. Meanwhile, at the NXT007 concentrations of 10 and 100 μg/mL, the TG capacity of HA plasma was not meaningfully altered by the coexistence of emicizumab.

Figure 1.

In vitro thrombin generation (TG) activity in the coexistence of emicizumab and NXT007. Effect of NXT007 in the coexistence of emicizumab on the peak height of TG in HA plasma. The reaction was triggered by TF (A) or FXIa (B). Data are expressed as mean ± SD (n = 3). HA, hemophilia A; TF, tissue factor; FXIa, activated factor XI.

3.2. In vitro binding properties of NXT007: competition between emicizumab and NXT007 for the antigen-binding epitopes

To clarify whether emicizumab and NXT007 compete for antigen binding, we performed BLI analysis, an optical technology that can measure real-time kinetics of protein-protein interactions [14]. Here, we utilized this method to clarify whether unintended NXT007-FIXa-emicizumab or NXT007-FX-emicizumab ternary complexes are formed. After emicizumab was first applied and bound to either biotinylated (bio-)FIXa or bio-FX immobilized on sensor chips, antibodies recognizing different epitopes from emicizumab (XB12 for FIXa, binder for the catalytic domain in FIXa; SB04 for FX, binder for the activation peptide in FX), NXT007, or emicizumab was sequentially applied (Figure 2A). As a result, XB12 or SB04 increased the binding signal, whereas neither NXT007 nor emicizumab did (Figure 2B and C). Similar results were observed when the experiment was conducted with NXT007 applied first. These results suggest that emicizumab and NXT007 recognize competing sites on both FIX(a) and FX. This indicates that unexpectedly larger complexes with tetrameric or higher stoichiometry cannot form from FIX(a), FX, emicizumab, and NXT007. It also indicates that emicizumab and NXT007 form distinct antigen-antibody complexes.

Figure 2.

In vitro competition analysis of emicizumab and NXT007. (A) Schematic diagram of in vitro competition analysis in BLI. Bio-FIXa was immobilized on the sensor tip as an example. (B) Sensorgram of BLI analysis for competition against FIXa. Bio-FIXa was immobilized on the sensor tip and emicizumab (upper panel), NXT007 (middle panel), or TBS (lower panel) was administered, followed by the addition of emicizumab (blue line), NXT007 (red line), or anti-FIXa CD antibody (XB12, purple line). (C) Sensorgram of BLI analysis for competition against FX. Bio-FX was immobilized on the sensor tip and emicizumab (upper panel), NXT007 (middle panel) or TBS (lower panel) were administered, followed by the addition of emicizumab (blue line), NXT007 (red line) or anti-FX AP antibody (SB04, yellow line). Ab, antibody; AP, activation peptide; Bio, biotinylated; BLI, biolayer interferometry; CD, catalytic domain; FIXa, activated factor IX; FX, factor X; TBS, Tris-buffered saline.

To support the BLI results, we then used immunoblotting analysis to clarify whether the epitopes recognized by NXT007 in FIX(a) and FX had been altered compared to those recognized by emicizumab as a result of the engineering process—light chain alteration and additional amino acid mutations (Figure 3A). Individual EGF1 or 2 domains of FIX(a) or FX, conjugated with Fc, were recombinantly prepared and used alongside full length pdFIXa and pdFX for epitope analysis, as previously described for emicizumab [15] (Figure 3B). The immunoblotting analysis revealed that NXT007 anti-FIX(a) monospecific antibody recognized pdFIXa, rFIX EGF1/2-Fc, and rFIX EGF1-Fc conjugates, whereas it did not recognize rFIX EGF2-Fc conjugate (Figure 3C). Similarly, NXT007 anti-FX monospecific antibody recognized pdFX, rFX EGF1/2-Fc, and rFX EGF2-Fc conjugates, whereas it did not recognize rFX EGF1-Fc conjugate (Figure 3D). These results suggest that NXT007 anti-FIX(a) arm and anti-FX arm recognized epitopes in the EGF1 domain of FIX(a) and EGF2 domain of FX, respectively. As expected, the epitopes were located in the same domains as those recognized by emicizumab. Based on the epitope similarity, we also evaluated the comparability of the mechanism of action between emicizumab and NXT007. We prepared the NXT007 anti-FIX(a) monospecific one-armed antibody and anti-FX monospecific one-armed antibody and confirmed their specific binding to rFIX/pdFIXa and pdFX/pdFXa, respectively (Figure 4A). In an enzymatic assay, NXT007 dose-dependently enhanced FIXa-catalyzed FX activation, whereas this reaction was not observed when NXT007 anti-FIX(a) monospecific one-armed antibody, anti-FX monospecific one-armed antibody, or a mixture of both antibodies was added (Figure 4B). This suggests a comparable mechanism of action between emicizumab and NXT007; namely, BsAb-dependent bridging of FIXa and FX through recognition of epitopes in the same domains in FIX(a) and FX.

Figure 3.

Epitope analyses of NXT007. (A) Schematic images of the molecular structures of emicizumab and NXT007 (B) FIX(a)- and FX-related proteins used for electrophoresis. (C) The electrophoretic pattern of FIX(a)-related proteins stained by CBB in a nonreduced condition and the results of immunoblotting with the anti-FIX(a) arm of NXT007. (D) The electrophoretic pattern of FX-related proteins stained by CBB in a nonreduced condition and the results of immunoblotting with the anti-FX arm of NXT007. (C, D) Bands >100 kDa are considered to be the aggregated multimers. CBB, Coomassie brilliant blue; EGF, epidermal growth factor; FIX, factor IX; FIXa, activated factor IX; FX, factor X; pd, plasma-derived; r, recombinant.

Figure 4.

In vitro cofactor activity mechanism of NXT007. (A) Sensorgram of SPR analysis. NXT007, anti-FIX(a), or anti-FX monospecific one-armed antibody (BsAb containing one anti-FIX(a) or anti-FX arm, and one anti-KLH [keyhole limpet hemocyanin; nonmammalian origin protein] arm) were captured on the sensor tip, and the binding to rFIX, pdFIXa, pdFX, or pdFXa was detected. X-axis represents time (s) and Y-axis represents response unit. (B) Effect of NXT007, anti-FIX(a) or anti-FX monospecific one-armed antibodies on FIXa-catalyzed FX activation in an enzymatic assay using purified coagulation factors and phosphatidylserine-containing phospholipid. FIX, factor FIX; FIXa, activated factor IX; FX, factor X; FXa, activated factor X; pd, plasma-derived; r, recombinant; SPR, surface plasmon resonance.

3.3. Simulated FIX-BsAb-FX ternary complex formation in plasma in the coexistence of emicizumab and NXT007 to interpret the TG results

The BLI analysis showed that emicizumab and NXT007 form distinct antigen-antibody complexes, indicating that plasma coagulation potential represents the sum of the contributions of both BsAbs. Therefore, we anticipated that TG capacity in their coexistence could be predicted from the respective plasma levels of FIX-BsAb-FX ternary complexes and the respective cofactor activity per ternary complex.

First, we carried out K_D-based simulation of each plasma FIX-BsAb-FX ternary complex levels in the coexistence of NXT007 and emicizumab. We previously performed such K_D-based simulations with each BsAb individually and revealed that the relationship of plasma concentration between BsAb and simulated FIX-BsAb-FX ternary complex was represented by a bell-shaped curve [10,15]. In addition, this bell-shaped relationship was also observed between BsAb levels and peak TG activities [10,15]. These previous results suggested that plasma FIX-BsAb-FX ternary complex levels correlate with topical ternary complex levels on the activated platelet surface. In this study, simulations were performed in the coexistence of both drugs using concentrations of 10, 50, or 300 μg/mL for emicizumab and a varying range for NXT007, corresponding to the conditions employed in the TG assays (Figure 1A, B). The simulations showed that the coexistence of emicizumab at 10 or 50 μg/mL had little impact on the level of FIX-NXT007-FX complexes compared to NXT007-only conditions (Figure 5A). Even in the coexistence of emicizumab at 300 μg/mL, which far exceeded the therapeutic plasma concentration, FIX-NXT007-FX formation still occurred. The peak levels were approximately 2-fold higher than FIX-emicizumab-FX levels, although the levels remained lower than those in conditions without emicizumab. On the other hand, at any concentration of emicizumab examined, the level of FIX-emicizumab-FX complexes showed a clear reduction with the addition of >10 μg/mL NXT007 (Figure 5B). These results suggest that the coexistence of a therapeutic concentration of emicizumab does not substantially interfere with the formation of FIX-NXT007-FX complexes, whereas NXT007 at the target concentration range (10-100 μg/mL) diminishes FIX-emicizumab-FX complex formation.

Figure 5.

Simulation of FIX-BsAb-FX ternary complex formation in plasma in the coexistence of emicizumab and NXT007. Effect of NXT007 on the coexistence of emicizumab in the simulated FIX-BsAb-FX ternary complex formation in plasma. The K_D-based simulation of the FIX-BsAb-FX complex was carried out based on the competitive binding of emicizumab and NXT007 with each antigen. The levels of FIX-NXT007-FX (A) and FIX-emicizumab-FX (B) are shown. (C) The theoretical cofactor activity in the presence of both emicizumab and NXT007 relative to emicizumab at 50 μg/mL (therapeutic concentration), based on FIX-emicizumab-FX and FIX-NXT007-FX levels and their in vitro conversion factors. Black dotted line represents value of emicizumab at 50 μg/mL. BsAb, bispecific antibody; FIX, factor IX; FX, factor X; K_D, dissociation constant.

Next, we examined the validity of this simulation using the TG results. Based on the results of the previous studies in the presence of either BsAb individually [10,15], the TG activity per simulated FIX-emicizumab-FX ternary complex at 50 μg/mL emicizumab was approximately 1.44-fold lower than that per FIX-NXT007-FX ternary complex at 1.7 μg/mL NXT007, which exerted the equivalent TG activity (Supplementary Figure 1A). Reflecting this activity difference per FIX-BsAb-FX complex, we calculated the theoretical cofactor activity in the coexistence relative to single presence of emicizumab at 50 μg/mL as follows: [{plasma FIX-emicizumab-FX concentration} × 1 + {plasma FIX-NXT007-FX concentration} × 1.44] / 0.905 nM (plasma FIX-emicizumab-FX concentration at 50 μg/mL of emicizumab) (Figure 5C). As a result, the calculated theoretical cofactor activities in the coexistence aligned with the in vitro TG results regarding the relative order of the activities across conditions evaluated (Figure 1A, B). This concordance, in turn, logically strengthened the insights from the TG results suggesting that the plasma coagulation potentials were neither abnormally enhanced nor impaired.

4. Discussion

As new treatment options for HA are being developed [17,18], understanding potential interactions when multiple therapies temporarily coexist during treatment transitions is crucial for persons with HA. This study demonstrated the in vitro interactive effect of emicizumab and NXT007 in their coexistence.

A key question is whether the cofactor activity is abnormally enhanced or impaired when emicizumab and NXT007 coexist. Regarding the abnormal enhancement, one of the possible risk scenarios is that NXT007 and emicizumab form unintended complexes with high coagulation enhancing activity only in their coexistence. Therefore, to address whether abnormal enhancement occurs, we combined the antigen competition analysis using BLI and the TG assay. We confirmed that emicizumab and NXT007 could not simultaneously bind to the same antigens, indicating that only ternary complexes or lower stoichiometry can be formed from FIX(a) and FX(a) (Figure 2B, C). In line with this logic, the TG capacity in the coexistence of both BsAbs was not abnormally enhanced (Figure 1A, B). Generally, when different antibodies recognize different epitopes on the same antigen, a drug-target-drug complex (DTDC) may form depending on the epitope or binding strength [19]. In persons with HA, formation of unexpected DTDCs with other drugs in plasma is undesirable for both safety and efficacy reasons (Supplementary Figure 2). The present study clearly indicated that the combination of emicizumab and NXT007 does not theoretically form such undesirable DTDCs, and DTDC-related risks can be ignored.

Regarding the abnormal impairment, the next possible risk scenario based on the binding profiles of NXT007 and emicizumab is that they compete with each other for binding to antigens, drastically diminishing the formation of their active complexes when both are present compared with either agent alone. To address whether abnormal impairment occurs, we combined the K_D-based simulation of each plasma FIX-BsAb-FX ternary complex levels and the TG results in the coexistence. The BLI results, which showed that emicizumab and NXT007 form distinct antigen-antibody complexes (Figure 2B, C), prompted us to interpret the coagulation potentials in the coexistence as the sum of the contributions of respective FIX-BsAb-FX ternary complexes. These contributions were calculated as the product of respective ternary complex levels and the corresponding conversion factors from FIX-BsAb-FX ternary complex to the activity driving coagulation. A previous report showed that emicizumab at 50 μg/mL and NXT007 at 1.7 μg/mL showed comparable peak thrombin values [10]. Focusing on this peak thrombin value, the conversion factor from FIX-NXT007-FX ternary complex to equivalent FVIII activity in TG capacity slightly increased compared to that from FIX-emicizumab-FX (approximately 1.44-fold: 22.2 vs 15.5; Supplementary Figure 1B). Indeed, the calculated coagulation potentials in the coexistence correlated with the ranking of TG capacity under evaluated conditions, showing no abnormal activity impairment (Figures 1A, B and 5C). This consistency, in turn, supported the approximate estimation of the coagulation potential utilizing the FIX-BsAb-FX ternary complex simulation as well as these ideal TG results.

For the analysis, we used the value 1.44-fold as a difference of TG-based conversion factor. We also confirmed the value using the turnover rate (k_cat) of FIXa-catalyzed FX activation in the presence of NXT007 based on an enzymatic assay. When compared chronologically with the previous report on emicizumab [15], consistent results were observed for NXT007 compared with emicizumab (approximately 2.1-fold: 6.12 vs 2.88/min; Supplementary Figure 1B for NXT007, Supplementary Table in the previous report [15] for emicizumab), considering the day-to-day variability in k_cat measurements (eg, k_cat driven by FVIIIa: 188/min for this study vs 126/min for the previous study; Supplementary Figure 1B, Supplementary Table in the previous report [15]). This consistency supported the use of the TG-based conversion factor as one of the calculation tools. These results suggest that efficient FIX-NXT007-FX complex formation, while maintaining enzymatic turnover efficiency, was the main mechanism for enhanced cofactor activity of NXT007 compared with emicizumab. Enhanced binding activity may negatively affect the turnover rate due to the impact on the substrate activation cycle. However, the turnover rate of NXT007 was maintained, possibly due to the optimization focusing on association rate (k_a) rather than dissociation rate (k_d) against FX [10].

Our study focuses on using plasma FIX-BsAb-FX ternary complex calculated from K_D values as an indicator of FIXa-BsAb-FX on lipid surfaces. We utilized the antibody-immobilized SPR method for K_D measurement, rather than direct measurement in solution, such as the isothermal titration calorimetry method or measurement of the complex on lipid surfaces. The methodology used for K_D measurement may affect the results. Other research teams reported the mathematical modeling and proposed efficient binding of emicizumab to lipid-bound FX and lipid-bound FIXa [20,21]. Despite the limitations of our K_D measurement methodology, we confirmed that the concentration of emicizumab or NXT007 that produced peak FIX-emicizumab-FX or FIX-NXT007-FX levels, respectively, correlated well with their peak TG activities [10,15]. Therefore, using the K_D values and the calculated plasma FIX-BsAb-FX levels as indicators of topical FIXa-BsAb-FX appears reasonable.

Emicizumab is administered at dosing regimens typically reaching a concentration ∼50 μg/mL at steady state in plasma from persons with HA [5]. Therefore, when considering transitioning from emicizumab to NXT007 in persons with HA, NXT007 may be administered in the presence of emicizumab at ∼50 μg/mL. Here, we focused on therapeutic situations to reexamine the TG results and the simulated FIX-BsAb-FX ternary complex levels. In TG assays with emicizumab at 50 μg/mL and varying concentrations of NXT007, an add-on effect was observed at low subtherapeutic NXT007 doses of approximately 0.1 to 1 μg/mL, while activity equivalent to the NXT007-only condition was observed at doses of 10 to 100 μg/mL (Figure 1A, B). These TG results ideally suggest the efficacy and safety of seamless switching from emicizumab to NXT007. The plasma FIX-NXT007-FX complex level was shown to be minimally affected at therapeutic emicizumab concentrations by simulating conditions with epitope competition between emicizumab and NXT007 (Figure 5A). In contrast, adding NXT007 at concentrations of 10 to 100 μg/mL decreased the amount of FIX-emicizumab-FX complexes (Figure 5B), consistent with the observation that the add-on effect of emicizumab was not clearly shown in the TG assays (Figure 1A, B). NXT007 has higher antigen-binding affinity than emicizumab, especially for FX [10], resulting in improved FIX-BsAb-FX formation efficiency compared to emicizumab. Therefore, at NXT007 concentrations of 10 to 100 μg/mL, when relatively reduced free FIX or FX condition, NXT007 outcompetes emicizumab and their coexistence would not provide the additive effect.

In the present study, the cofactor activity evaluations were mainly based on the peak thrombin value in TG assessment, one of the in vitro closed assay systems. Comprehensive coagulation tests can produce different outputs depending on the assay system components, trigger conditions, and the parameter selections. In our understanding, there is no single definitive evaluation system that totally reflects in vivo hemostasis. However, we believe that the TG capacity can serve as a relative guidepost for cofactor activity comparison, considering that the TG peak height well reflected the nonclinical hemostatic potential in the case of emicizumab [10].

This study focused on the interactive effect of emicizumab and NXT007. With the growing number of clinically used treatment options, persons with HA would also have a chance to experience a coexistence in blood with NXT007 and other drugs, such as FVIII-replacement therapies, bypassing agents, and rebalancing agents [17,18]. Due to the difference of mechanisms of action, these agents would theoretically show an additive or synergistic effect on coagulation potential. A previous report showed concomitant addition of rFVIII, rFVIIa, and activated prothrombin complex concentrate increased coagulation potentials in vitro in the presence of NXT007 [22]. The actual impact on the level of coagulation potential and the risk of hypercoagulation in that combination should be carefully investigated in the clinical setting.

In conclusion, the present study provides supportive evidence for transitioning from emicizumab to NXT007 without emicizumab washout while maintaining enhanced but not exceeding the targeted coagulation potential range in plasma. Further clinical evaluations of NXT007 in the context of a prior emicizumab administration regimen will investigate the actual impact of these results in clinical settings. The first clinical study evaluating the effect of switching from emicizumab to NXT007 is currently ongoing. [jRCT2080224835].