Model-informed drug development of Mim8 – a next-generation bispecific antibody for treatment of haemophilia A

Abstract

Background

Haemophilia A is a congenital bleeding disorder caused by deficiency of clotting factor VIII (FVIII). Mim8 (denecimig) is a next-generation activated FVIII mimetic bispecific antibody under phase 3 investigation for subcutaneous management of haemophilia A with or without FVIII inhibitors. Mim8 uses tiered dosing, which is a novel dosing approach in haemophilia where patients receive a fixed dose according to which body weight range they fall into and their chosen dosing frequency, reducing the need for dose calculations and potentially reducing treatment wastage.

Objectives

This study aimed to establish the early population pharmacokinetics (PK) and PK/pharmacodynamic (PD) properties of Mim8 using both biomarker and clinical endpoint data. Models were used to inform drug-development decisions using trial simulations and transform the dosing strategy in haemophilia to a novel simplified treatment option using tiered dosing in paediatric and adult populations.

Patients/Methods

Data from 129 participants exposed to Mim8 in the phase 1 PK trial (NCT0512747) and the FRONTIER1 trial (NCT04204408) were used to develop population PK and PK/PD models for thrombin generation and treated bleeds. The models were used for trial simulations to optimise dosing regimens evaluating Mim8 exposure, biomarker and clinical endpoint targets on a population level (paediatric and adolescent/adult) and on a typical subject level for the range of body weight (BW) in the trial. Clinical endpoint targets were set to: (1) at least 70 % of patients without treated bleeds in a 6-month trial on a population level and >95 % reduction in relative bleed risk on a typical subject level across all weight ranges and dosing frequencies, and (2) sustained peak Mim8 exposure below highest exposure observed in phase 1/2 trials.

Results

The population PK of Mim8 were best described with a structural two-compartment model with baseline BW as the most significant covariate for Mim8 exposure. Both peak thrombin response and Mim8 effect on bleed risk were best described with a direct link PK/PD model with an E_max response. Model-estimated EC₅₀ was 2.03 µg/mL for peak thrombin and 0.07 µg/mL for bleed risk. Using trial simulations, tiered dosing in three weight ranges (<15, 15–<45, and ≥45 kg) was identified to reach the clinical endpoint targets and maximum desired exposure for all three dosing frequencies: once every week, once every 2 weeks, and once every month. This was done with an initial one-time loading dose establishing steady-state conditions within the first day, followed by maintenance doses at the desired dosing frequency. Loading and maintenance doses for the nine combinations of weight ranges and dosing frequencies were covered by five dose presentations. Ongoing phase 3 trials have been based on the dosing strategy presented here.

Conclusions

Model-informed drug development was used to design a novel treatment paradigm for patients living with haemophilia A by utilising a tiered-dosing strategy with Mim8, to be evaluated in phase 3 trials.