Ruben Faelens1, Belén Valenzuela2, Martine Neyens1, Anne Gaëlle Dosne1, Juan José Pérez Ruixo2
1Johnson & Johnson, 2Johnson & Johnson
Background: Generalized myasthenia gravis (gMG) is a rare, chronic, neuromuscular disease caused by pathogenic autoantibodies that impair or prevent muscle contraction. Circulating autoantibodies targeting the acetylcholine receptor (AChR) are present in 85% of cases, the muscle-specific tyrosine kinase (MuSK) in 7-8% of cases, while ~7% to 8% of patients have neither anti-AChR nor anti-MuSK antibodies (“seronegative” gMG). Nipocalimab is a fully human immunoglobulin G (IgG) 1 monoclonal antibody (mAb) that binds to FcRn with high affinity and blocks binding of endogenous IgG to FcRn, inhibiting IgG recycling, decreasing IgG half-life, and thus rapidly reducing the circulating levels of IgG, including pathogenic IgG autoantibodies. Nipocalimab demonstrated efficacy, favorable safety and tolerability, and dose-dependent IgG reduction in a Phase 2 study (M281-004 study), and in the pivotal Phase 3 study (M281-011 study) in patients with gMG. A comprehensive semi-mechanistic population model was developed integrating the nipocalimab serum concentrations, FcRn occupancy and IgG reduction data from Ph1 clinical studies in healthy participants and M281-004 study, accounting for the relationship between total serum IgG reduction and MG-ADL scores reduction in patients with gMG. The model fit the Phase 1/Phase 2 data well. [1] Methods: The predictive performance of the model was evaluated by predicting the clinical efficacy outcomes of the Phase 3 study and comparing with the observed outcomes. A total of 2,000 clinical study replicates with the same sample size as the Phase 3 study were simulated. The individual MG-ADL observations and predictions (with residual error) were used to derive the selected clinical efficacy endpoints for placebo and nipocalimab arm. The results were presented as the median (95%CI) for each arm and compared with the observed value in the Phase 3 clinical study. The observed and predicted between-group difference (drug effect) (95%CI) was also determined. The 5 selected clinical efficacy endpoints investigated were: 1.Endpoint 1: Average MG-ADL reduction from baseline at Week 22-24. 2.Endpoint 2: Proportion of responders over Weeks 22-24. Percentage of participants whose average improvement in MG-ADL total score over Weeks 22-24 of the double-blind placebo-controlled phase is at least a 2-point improvement compared to baseline. 3.Endpoint 3: Proportion of responders within the first 2 weeks of dosing. Percentage of participants with improvement in MG-ADL total score =2 points at Week 1 and/or Week 2 of the double-blind placebo-controlled phase. 4.Endpoint 4: Sustained response. Percentage of participants with improvement in MG-ADL total score =2 points at Week 4 through Week 24 of the double-blind placebo-controlled phase with no more than 2 non-consecutive excursions allowed between Week 6 through Week 23 (excursion defined as loss of improvement in MG-ADL score =2 points from baseline). 5.Endpoint 5: At least 50% symptom improvement. Proportion of subjects with 50% or greater symptom improvement in MG-ADL scores from baseline: Percentage of participants whose average improvement in MG-ADL total score over Weeks 22, 23, and 24 of the double-blind, placebo-controlled phase is at least a 50% improvement from baseline. Results: Endpoints 1, 2, 4, and 5 were well predicted by the model for both placebo, nipocalimab, as well as the predicted per-study drug effect, as the observed endpoints from the Phase 3 study were within the 95% CI of the clinical trial simulations from the Phase 2 model. In particular, the average MG-ADL reduction from baseline at Week 22-24 was well predicted for both placebo (-3.22 observed vs -2.86 predicted, with 95% CI between -1.87 and -3.88), treatment (-4.98 observed vs -4.86 predicted, with 95% CI between -3.96 and -5.76), and the difference between placebo and treatment (-1.75 observed vs -2.01 predicted, with 95% CI between -0.67 to -3.35). The proportion of early responders (endpoint 3) in the treatment arm was overpredicted by the model. Conclusion: A comprehensive nipocalimab semi-mechanistic population model can integrate the pharmacokinetics, biomarker and efficacy endpoints from early clinical studies in patients with gMG and accurately and precisely predict almost all clinical efficacy endpoints of a Phase 3 study. These results underscore the potential of pharmacometric approaches as a tool for optimizing dosing regimens and guiding clinical study design.
[1]: Nipocalimab Dose Selection for A Phase 3 Study in Adult Patients with Generalized Myasthenia Gravis, Perez Ruixo et al, PAGE 31 (2023) Abstr 10385 [www.page-meeting.org/?abstract=10385]
Reference: PAGE 33 (2025) Abstr 11427 [www.page-meeting.org/?abstract=11427]
Poster: Drug/Disease Modelling - CNS