A translational quantitative systems pharmacology modeling framework for FcRn antagonist therapy for myasthenia gravis

Objectives: Myasthenia gravis (MG) is an autoimmune disorder with substantial unmet medical needs. MG is often caused by IgG autoantibodies that attack the nicotinic acetylcholine receptors at the neuromuscular junctions, resulting in direct pathogenic effects. Efgartigimod is a newly approved drug by the FDA for MG treatment. Efgartigimod is a human IgG1 Fc-fragment that binds to FcRn with high affinity, reducing total IgG including pathogenic autoantibodies by inhibiting their FcRn-mediated recycling. Efgartigimod clinical data have shown a reduction in IgG concordant with clinically meaningful improvements measured by the QMG and MGADL scores (1,2), suggesting the pathogenic autoantibody is a disease related biomarker in MG. To facilitate the discovery and development of autoantibody targeting MG treatments, it is crucial to understand the relationship between pharmacokinetics (PK), pharmacodynamics (PD), and clinical scores. In this work, we used the available clinical data to develop a quantitative systems pharmacology (QSP) model to study PK, autoantibody levels, and MG scores relationships in efgartigimod treated MG patients.Methods: First, a minimal physiologically based PK (mPBPK) model was utilized to describe PK and PD. The phase 2 PK/PD data were used for the mPBPK model fitting. Second, the relationship between the autoantibody reductions and MG scores was described by a linear model connecting the autoantibody and clinical scores with an additive term to account for the time-dependent placebo effect. The phase 3 data were used for the second model fitting. Lastly, the integrated QSP model was used to assess the model performance against the clinical data. Results: Accounting for the competitive inhibition of IgG recycling by incorporating the FcRn binding kinetics, the mPBPK model successfully described the longitudinal phase 2 PK/PD observations; ~50% autoantibody reduction during the 4-week treatment course and the recovery to baseline over ~8 weeks following the treatment period. Following adjustment for the placebo effect, the relationship between autoantibodies and clinical MG scores was characterized by a time-independent linear model (MGADL r=0.96, QMG r=0.98), highlighting a strong link between autoantibodies and clinical scores after a therapy targeting autoantibodies. A global sensitivity analysis showed that efgartigimod binding on-rate (to FcRn), degradation, and recycling rates are the most influential parameters on the autoantibody levels. Conclusions: A modeling framework was developed for efgartigimod treatment of MG patients. By delineating the longitudinal relationship between autoantibodies and MG scores and accounting for placebo effects, the developed framework will serve as a quantitative means to benchmark the current autoantibody-targeting MG treatments and provide a translational basis to accelerate the drug discovery and development of next generation of therapies targeting autoantibodies for MG patients.Citations: [1] J. Howard et al., Neurology, 92: e2661-73 (2019). [2] J. Howard et al., Lancet Neurol, 20: 526–36 (2021).