Population pharmacokinetic-target engagement analysis of ocular and systemic data for intravitreal BI 764524 in patients with diabetic macular ischemia from the HORNBILL study
Qing Xi Ooi1, Anyue Yin1, Bartlomiej Krawczyk2, Martin Gliem3, David Brown4, Chirag Jhaveri5,6, Victor H. Gonzalez7, Charles Wykoff8, Khaled Nassar2, Stefan Wölke9, Gudrun Simons9, Michel Wagner9, Sobha Sivaprasad10, Quan Dong Nguyen11, Ronald Niebecker9
1Pharmetheus AB, 2Boehringer Ingelheim Pharma GmbH & Co. KG, Ingelheim am, 3Boehringer Ingelheim International GmbH, Ingelheim am, 4Retina Consultants of Texas, Retina Consultants of America, 5Retina Consultants of Austin, 6Austin Research Center for Retina and Dell Medical School, 7Valley Retina Institute, 8Retina Consultants of Texas and Blanton Eye Institute, Houston Methodist Hospital, 9Boehringer Ingelheim Pharma GmbH & Co. KG, 10NIHR Biomedical Research Centre for Ophthalmology, Moorfields Eye Hospital, 11Byers Eye Institute, Stanford University
Introduction/Objectives: One of the anatomical changes of diabetic retinopathy (DR) is retinal non-perfusion, centrally manifesting as diabetic macular ischemia (DMI). BI 764524, a novel monoclonal antibody administered intravitreally that binds and neutralizes Semaphorin 3A (Sema3A), showed good tolerability, safety and potential for disease-modifying effects on DMI in proliferative DR [1]. The aim of this analysis was to characterise the pharmacokinetics (PK) and Sema3A engagement properties of intravitreal (IVT) BI 764524 to support selection of dosing regimens for the Phase IIb study. Methods: HORNBILL was a two-part, Phase I/IIa study in patients with DMI. Part 1 was non-randomised, open label, with single rising doses (SRD) (BI 764524 0.5, 1.0 or 2.5 mg). Part 2 was randomised, masked, sham-controlled, with multiple doses (MD) (sham or BI 764524 2.5 mg on Day 1 and Weeks 4 and 8). Optional aqueous humour (AH) samples were collected in the MD part (Day 1 and Weeks 8, 16 and 22) in addition to serum samples from the SRD part (Days 1 and 4 and Weeks 1, 2, 4, 8 and 14) and MD part (Day 1 and Weeks 4, 8, 12, 16 and 22). The PK-target engagement (TE) analysis was performed using NONMEM version 7.5. The joint-sequential population PK parameters and data (PPPD) approach was used to consider the AH data preferentially before including the serum data. The overall model structure was based on the retina–vitreous humour (VH)–AH model for ranibizumab developed by Hutton-Smith et al. [2]. Potential covariate-parameter relationships were evaluated using the stepwise covariate model (SCM) building procedure with adaptive scope reduction. Results: The analysis data set included: (a) total BI 764524 concentrations in AH (number of observations, n=15; number of subjects, N=7); (b) total Sema3A concentrations in AH (n=36, N=10); (c) total BI 764524 concentrations in serum (n=214, N=32); (d) total Sema3A concentrations in serum (n=260, N=42). The final PKTE model consists of three connected one-compartment models describing the PK and TE of BI 764524 in vitreous humour (VH), AH and serum, respectively. The model describes (a) the associations and dissociations of BI 764524 with one or two Sema3A molecules in VH, AH and serum and (b) the elimination of free BI 764524 (D), free Sema3A (S), BI 764524–Sema3A complexes (DS), and Sema3A–BI 764524–Sema3A complexes (SDS) from VH to AH and from AH to serum. The BI 764524–Sema3A engagement properties were described by a full target-mediated drug disposition (TMDD) model with unidentifiable parameters fixed to relevant values reported in literature and in the non-clinical report. No covariate-parameter relationship was included in the model. The final PKTE model assumes that the turnover of D, S, DS and SDS in AH are negligible; only 1:1 (DS) and 2:1 (SDS) bindings are possible between BI 764524 and Sema3A; and binding of one Sema3A to BI 764524 does not alter the binding affinity of the second Sema3A molecule. The goodness-of-fit plots and the visual predictive check plots showed that the final model provided an adequate description of the observed data. Conclusions: A VH-AH-serum PKTE model was developed to characterise the PK and Sema3A engagement properties of IVT BI 764524 based on AH and serum total BI 764524 and total Sema3A concentration data in patients with DMI from the HORNBILL study. The final model provided an adequate description of the observed data and may be used for simulations to support selection of dosing regimens for the Phase IIb study.
[1] Nguyen QD et al. Presented at ARVO 2024. Presentation 959. [2] Hutton-Smith LA et al. Model Mol Pharm 2018;15:2770–2784.