Target engagement analysis of belimumab in patients with systemic lupus erythematosus (SLE) and primary Sjogren’s syndrome (pSS)
Richard Dimelow PhD (1), James WT Yates PhD (2), Eric Salgado PhD (3), Enrica Mezzalana PhD (3)
(1) Clinical Pharmacology Modelling and Simulation, GSK, Stevenage, UK, (2) DMPK Modelling, GSK, Stevenage, UK, (3) Pharmetheus AB, Uppsala Science Park, Uppsala, Sweden
Introduction: Elevated B-lymphocyte stimulator (BLyS) levels are found in B-cell-mediated autoimmune diseases such as SLE and pSS[1,2]. Belimumab (BEL) inhibits BLyS and provides clinical benefit to patients with SLE[1,3]. BEL has also been tested in pSS but with less clear efficacy[4]. Intravenous (IV) BEL 10 mg/kg every 4 weeks and subcutaneous (SC) 200 mg every week are approved for SLE.
Objectives: To develop a target-mediated drug disposition (TMDD) model for BEL based on total BEL pharmacokinetic (PK) and total BLyS data from the pSS (201842)[2] and SLE (205646)[5] studies and to characterise BEL target engagement (TE) in patients with SLE for the approved doses.
Methods: The TMDD model development used data from the SC BEL studies 201842 and 205646. Total drug and free and total BLyS concentrations were measured over time. Free BLyS was excluded as its measurements were considered an overestimate of their true value due to redistribution of BLyS in the binding assay[6], consistent with a visual predictive check comparing actual and simulated observations. Thus, the model was fitted to total drug and total BLyS data only.
A two-compartment PK model with first-order absorption and with a TMDD model component to describe BEL binding to BLyS in the central compartment was developed based on a previous population PK model for SC BEL[7]. Body weight (WT) was included as a mechanistic covariate on all clearance (CL) and volume of distribution (V) parameters, with fixed allometric exponents of 0.75 and 1, respectively. The BLyS synthesis rate (ksyn) was defined as ksyn=BASE·kdeg, where the baseline BLyS level (BASE) is assumed equal for SLE and pSS, consistent with observed data, and kdeg is the first-order degradation rate constant of free BLyS. The degradation rate of the BEL–BLyS complex (kint) was also assumed equal for SLE and pSS.
The model was used to simulate BEL TE, defined as percentage of free BLyS depletion from baseline, for 1, 4 and 10 mg/kg IV and 200 mg SC BEL dosing in a virtual population of 1000 patients. The simulated TE across a dosing interval at steady state (SS) was characterised by the minimum and average TE in the top 80% to 95% of the population, and by the percentage of the population having minimum and average TE above 80% to 95%.
In a post hoc analysis, the model was used to explore TE in the two Phase 3 SLE studies (BLISS-52 and BLISS-76) that measured efficacy for 1 and 10 mg/kg IV doses.
Results: The BEL TE model, using the quasi-SS approximation, was fitted to the data without free BLyS observations. PK estimates were: CL 0.193 L/day; total V 5.54 L; BASE 0.0513 nM; and kdeg 1.14/day, corresponding to a half-life of 14.6 h. The quasi-SS affinity constant estimate (kss 541 pM) was comparable with the in vitro reported value (~1500 pM).
Across the dosing interval at SS, only the approved 10 mg/kg IV and 200 mg SC doses achieved a minimum TE of >90% in at least 95% of the population, while 1 and 4 mg/kg IV doses were expected to achieve a minimum TE >90% in only 15.6% and 80.2% of the population, respectively. Similarly, an average TE of >95% in at least 95% of the population was expected for the approved 10 mg/kg IV and 200 mg SC doses, while only 5.3% and 67.7% of the population met this condition for the 1 and 4 mg/kg IV doses.
PK and BLyS levels derived from the BLISS-52 and BLISS-76 study data supported these findings. Across all patients from the 1 and 10 mg/kg IV doses, there was a positive correlation between efficacy and average TE at SS, with maximum efficacy achieved for average TE of 95%. Average TE at SS was generally >95% for the 10 mg/kg IV dose but <90% for the 1 mg/kg IV dose. The efficacy response rate per dose group was 46.2% (1 mg/kg) and 50.6% (10 mg/kg). Thus, these studies showed average TE at SS >95% is associated with maximum BEL efficacy, which is achieved for 10 mg/kg IV.
Conclusion: PK and total BLyS levels from SLE and pSS studies were combined in a TMDD model to characterise TE in response to BEL. The omission of free BLyS did not impact the model fit with the BEL–BLyS binding affinity estimate in line with in vitro measures. A high degree of TE was predicted for the approved doses, 10 mg/kg IV and 200 mg SC, compared with 1 and 4 mg/kg IV doses. Maximum efficacy was associated with an average TE at SS >95% and minimum TE at SS >90%; this was achieved in at least 95% of patients for the approved SLE dosing regimens. Funding: GSK.
References:
[1] Furie R et al. Arthritis Rheum 2011;63(12):3918–3930
[2] Mariette X et al. JCI Insight 2022;7(23):e163030
[3] Navarra SV et al. Lancet 2011;377:721–731
[4] Park HS et al. Ann Rheum Dis 2018;77:1417 (AB0518)
[5] Aranow C et al. Arthritis Rheumatol 2021;73 (suppl 9)
[6] Fairman D and Tang H. AAPS J 2023;25:23
[7] Struemper H et al. Clin Pharmacokinet 2018;57(6):717–728
