Confirming pharmacokinetic properties of satralizumab using a bodyweight tiered dosing regimen in LUMINESCE, a phase 3 study for generalized Myasthenia Gravis
Hanna Silber Baumann1, Sebastien Jolivet1, Ekaterina Gibiansky2, Jean-Christophe Genin1, Jill Smith3, Hajime Ito4, Sian Lennon-Chrimes3
1Roche Pharma Research and Early development, Pharmaceutical Sciences, Roche Innovation Center Basel, 2QuantPharm LLC, 3Roche Products Ltd, 4Chugai Pharmaceutical Co., Ltd
Objective: Satralizumab (SA) is a humanized IgG2 monoclonal recycling IL-6 receptor antagonist antibody. It has been approved for treatment of neuromyelitis optica spectrum disorder (NMOSD) and is being evaluated in new indications. The safety and efficacy of SA in patients with generalized Myasthenia Gravis (gMG) has been evaluated in the LUMINESCE trial [1]. A model-informed bodyweight (BW)-tiered dosing regimen was implemented based on the pharmacokinetic (PK) properties observed in NMOSD in order to maximize the target engagement (RO) across the body weight range. The objective of the current analysis was to confirm the SA PK properties using this regimen in the gMG population including development and impact of anti-drug antibodies (ADA) as well as rescue therapy on SA PK. Methods: LUMINESCE was a Phase III, randomized, double-blind, placebo controlled, multicenter study designed to evaluate the efficacy, safety, pharmacokinetics, and pharmacodynamics of SA compared with placebo in patients with gMG on stable background therapy. The study included a 24-week double-blind (DB) treatment period, followed by an open label extension (OLE). 188 patients were randomized in a 1:1 ratio to receive either SA, or placebo. Patients were eligible to receive rescue therapy (including intravenous immunoglobulin (IVIg) and plasma exchange (PE)) in case of a gMG exacerbation while administration with SA continued as scheduled. A BW-tiered dosing regimen was implemented with a SA dose of 120 mg for patient with BW =100 kg or 180 mg for BW > 100kg administered as a sub-cutaneous (SC) injection on weeks 0, 2, 4, and every 4 weeks (Q4W) thereafter. PK and ADA were sampled pre-dose at each visit during the DB and for the first 24 weeks of the OLE, and every 12 weeks thereafter. The PK analysis was based on the previously developed (legacy) popPK model [2]. This was a two-compartment model with first order subcutaneous absorption and parallel linear and Michaelis-Menten elimination. BW was identified as a covariate on linear clearance (CL), inter-compartmental clearance (Q), and volumes (VC, VP) with fixed allometric scaling coefficients (0.75 and 1, respectively). For patients with NMOSD CL was higher at times when ADA were detected, and SC bioavailability (Fsc) was lower in patients that had ADAs any time during the study. A logistic regression analysis was performed to investigate the factors influencing the appearance of ADA, including exposure, BW and background therapy. Results: 1606 PK and ADA samples from 168 subjects were included in the analysis. 147 (88%) patients received 120 mg and 21 (12%) patients received 180 mg as their initial SA dose. The median treatment duration was 44 weeks. ADA were detected in 67 (40%) patients. 2 patients received PE and 6 patients received IVIg while receiving SA treatment. The popPK analysis showed that the PK in gMG patients was similar to NMOSD for ADA negative subjects. In ADA positive subjects, the initial model underpredicted the SA concentrations, indicating that the impact of ADA was weaker. This finding was consistent with lower ADA titers and later onset in the gMG patients. The popPK model was adjusted to exclude the ADA effect on Fsc while keeping all other parameters fixed, which resulted in appropriate description of the gMG PK data. The impact of PE and IVIg was evaluated as covariates on CL, and both were found to be significant although the small sample size resulted in high uncertainty especially for PE. The effect of IVIg declined exponentially with time, which reflects the clearance of IVIg. Low exposure was found to be the main predictor for the appearance of ADA and is the combination of several effects including difference in BW between ADA positive and negative patients. The BW-tiered dosing regimen provided additional insights to the correlation between exposure, BW and risk of developing ADA, compared to the flat dose previously used. Median RO was predicted to be >95% across the dose interval for the entire BW range, regardless of ADA status. Conclusion: The final model was very similar to the legacy NMOSD model, but without the effect of ADA on SC bioavailability and it accounted for additional clearance of PE and IVIg treatment. The per protocol instruction to re-load SA following PE was confirmed to be appropriate. The predicted RO (>95%) confirmed the appropriateness of the selected BW-tiered dosing regimen in the gMG population. Presence of ADA is of no consequence for the clinical benefit of SA.
[1] Safety and efficacy of satralizumab in patients with generalised myasthenia gravis (LUMINESCE): a randomised, double-blind, multicentre, placebo-controlled phase 3 trial. Habib, Ali A et al. The Lancet Neurology, Volume 24, Issue 2, 117 - 127 [2] PAGE 29 (2021) Abstr 9665 [www.page-meeting.org/?abstract=9665]
