Model-based dose selection for the Phase 2b study of ianalumab in primary Sjögren's Syndrome
Irina Baltcheva (1), Martin Fink (1), Olivier Petricoul (2), Didier Renard (1)
(1) Pharmacometrics, Novartis AG, Basel, Switzerland; (2) Pharmacokinetic Sciences, Novartis Institute for Biomedical Research, Basel, Switzerland
Objectives:
Ianalumab is a novel human monoclonal antibody (mAb) that binds to the B-cell Activating Factor receptor (BAFFr). It has a dual mode of action in vivo: (i) rapid and profound B-cell depletion via antibody-dependent cell-mediated cytotoxicity (ADCC) and (ii) blockade of the BAFF-BAFFr signaling of tissue-residing activated B cells. A dose range finding Phase 2b study in primary Sjögren Syndrome (pSS) patients is currently ongoing (NCT02962895). The objective of the PKPD modeling presented here was to support the selection of ianalumab dose levels and regimens to be tested in the above Phase 2b study considering both mechanisms of action. A population PKPD model for circulating B cells based on clinical data had been established previously. To account for the second mode of action for which data was not available, a hypothesis-driven model of tissue receptor occupancy (RO) had to be developed. Both models were used to support dose selection and this model-based approach was approved by Health Authorities.
Methods:
We used a modeling approach to establish the dose-exposure-response relationship for both biomarkers related to the compound’s mode of action: 1) circulating B cells and 2) BAFF receptor occupancy (RO) by ianalumab in a disease-related tissue. A PKPD model for the first biomarker was available, based on data from Rheumatoid Arthritis (RA) and pSS patients. The population PK model had two compartments with linear clearance. B cell dynamics were represented by a turnover model with a peripheral compartment. Serum drug concentration was assumed to exert an indirect effect on circulating B cells by stimulating their death rate. No data was available for the second biomarker; we therefore used a model to predict RO in a hypothetical disease-related tissue. The RO model links the serum concentration to the unmeasured tissue RO using a simple analytical expression that represents the competitive binding between ianalumab and soluble BAFF (sBAFF) on BAFFr under quasi-steady state conditions [1].
Results:
Simulations based on the above PKPD models supported the selection of three dose levels to be tested in the Phase 2b study in pSS patients. The “low” dose is expected to provide exposures that just lead to depletion of B cells in the systemic circulation. The “high” dose was selected to approximately match the exposure in the proof-of-concept (PoC) trial that demonstrated efficacy in pSS patients [2]. The “medium” dose was chosen to provide a good spread of exposures between the “high” and the “low” doses. It is intended to provide evidence of increased efficacy due to targeting the BAFFr pathway, in addition to the expected clinical benefit due to complete depletion of circulating B cells. This dose is needed to describe a full dose-exposure-response model in pSS patients.
Conclusions:
We present a model-based approach to Phase 2b dose selection which includes data- and hypothesis-driven models and that was accepted by Health Authorities. We developed a tissue RO model in order to fill the gap between the available biomarkers and our understanding of the mode of action of ianalumab. The selected dose range aimed at providing a wide spread of drug exposures in order to mitigate the risk associated with the assumptions behind the tissue RO model. Such a range will enable characterization of the exposure-response relationship with clinical endpoints and thus inform the Phase 3 dose(s) to be tested in pSS patients.
References:
[1] Marathe A., et al, “Pharmacokinetics and Pharmacodynamics of Anti-BR3 Monoclonal Antibody in Mice”, Pharm Res. 2012, Nov; 29(11): 3180–3187.
[2] Doerner T., et al, “Safety and Efficacy of Single Dose VAY736 (anti-BAFF-R mAb) in Patients with Primary Sjögren’s Syndrome (pSS)”, 2016 ACR/ARHP Annual meeting, abstract number 3033.