Model-informed development of Mim8 - a novel bispecific antibody for the treatment of haemophilia A
Mads Kreilgaard (1), Mads Bjelke (1), Paula Persson (1)
(1) Novo Nordisk A/S, Denmark
Objectives:
Mim8 is a novel, next-generation FVIIIa mimetic bispecific antibody (biAb) in Phase 3 clinical development for subcutaneous treatment of haemophilia A with/without FVIII inhibitors1. Assisted by model-informed drug development strategy, clinical development has been frontloaded to reduce time from first human dose to initiation of pivotal phase 3 in less than two years. Furthermore, combined PKPD models using biomarkers from in vitro and ex vivo studies, non-clinical and literature data have helped transform the treatment paradigm in haemophilia from mg/kg dosing using vial and syringe into a simple weight-band dosing approach with few single-dose devices covering multiple dosing regimens in the ongoing phase 3 studies.
Methods:
Prior to clinical data, the human prediction model of Mim8 PK (unpublished) was based on a cynomolgus monkey model. The model was scaled to humans by allometric methods. Furthermore, two PKPD prediction models were developed:
- Peak thrombin generation (biomarker for haemostasis) in FVIII-neutralised plasma in vitro from healthy subjects, spiked with different Mim8 concentrations to compare PD to a comparable biAb within the same drug class (emicizumab)
- Repeated time-to-event model of treated bleeds in haemophilia patients, assuming a comparable bleed distribution as emicizumab2, but with higher potency as assessed by the peak thrombin biomarker
The models were used to evaluate potential of Mim8 as a true once monthly dosing solution with a low volume single injection and to assess population variability in Mim8 PK and PD based on either a mg/kg dosing solution or a simpler weight band approach.
The key clinical PK and PKPD models were continuously updated with clinical data from a phase 1/2 study in healthy subjects (SAD) and patients with haemophilia A (MAD), FRONTIER13, and a phase 1 formulation study4 (4882).
Iterative study simulations were performed to support phase 3 dose and weight-band selection in adult/adolescent and paediatric populations to ensure safety exposure cap as identified in FRONTIER1 and high reduction in relative bleed risk were covered for all extremes in the weight band groups across QM, Q2W and QW dosing intervals.
The models were developed by means of non-linear mixed effects modelling using NONMEM v. 7.35.
Results:
The plasma concentration-time profiles of Mim8 were best described with a structural 2-compartment model with first-order absorption and elimination rate for both the prediction model from monkeys and human clinical data. Random effects included IIV on CL, ka and F and a proportional residual error in the clinical model with the baseline body weight (BW) included as allometric exponents on systemic PK parameters and drug product strength as covariate on relative bioavailability.
The peak thrombin in vitro PD profiles for Mim8 and emicizumab were best described by a common sigmoidal Emax model with IIV on Emax, EC50 and hill, and drug as covariate on all structural parameters. Mim8 was shown to generate equivalent peak thrombin response as emicizumab with 15-fold lower plasma concentration and have a 7% higher Emax compared to emicizumab.
Observed first human dose profiles (PK and ex vivo peak thrombin) for Mim8 were well described by the prediction models. Early trial simulations supported once monthly dosing strategy with a low-volume single injection and suggested marginal increased variability in Mim8 exposure on a population level compared to mg/kg dosing for the MAD part of FRONTIER1 in patients.
Treated bleeds were adequately captured by a repeated time-to-event model with constant baseline hazard and estimated Mim8 IC50 of 0.07 µg/mL on bleed risk in alignment with expectation from the peak thrombin biomarker.
The clinical PKPD models were used to define dosing regiments for the phase 3 studies6,7 in patients from 5-140 kg BW with QM, Q2W and QW dosing intervals using a single loading dose followed by a maintenance dose in three weight band classes from <15, 15-<45 and >=45 kg. Study simulations ensured the average exposure cap was not exceeded and supported >97% relative reduction in bleed risk across all BW. The 2x9 matrix of all doses could be covered by five single dose devices
Conclusions: The model-informed drug development of Mim8 was shown to adequately design a novel treatment paradigm for patients living with haemophilia by utilising a weight-band dosing in ph3 studies. Furthermore, to enable early decision making and regulatory guidance to accelerate drug development
References:
[1] Østergaard H, Lund J, Greisen P , et al. A factor VIIIa-mimetic bispecific antibody, Mim8, ameliorates bleeding upon severe vascular challenge in hemophilia A mice. Blood. 2016; 138:1258-1268.
[2] Yoneyama K, Schmitt C, Kotani N, et al. A Pharmacometric Approach to Substitute for a Conventional Dose-Finding Study in Rare Diseases: Example of Phase III Dose Selection for Emicizumab in Hemophilia A. Clinical pharmacokinetics. 2018; 57:1123-1134.
[3] FRONTIER1: NCT04204408; EudraCT:2019-000465-20
[4] EudraCT: 2021-003182-35
[5] Beal SL et al. 1989-2014. NONMEM Users Guides. Icon Development Solutions, Ellicott City, Maryland, USA
[6] FRONTIER2: NCT05053139; EudraCT:2020-001048-24
[7] FRONTIER3: NCT05306418; EudraCT:2020-003467-26