Vincent Madelain1, Hélène Aerts1, Mélanie Leheup1, Virginia Parks1, Loubna Chadli1, Guillaume Das Dores2, Manuela Ingallinesi3, Hélène Tran3, Valérie Olivier3, Sylvain Fouliard1
1Translational Medicine, Institut de Recherche Servier Paris-Saclay, 2In vivo pharmacology activity pole, Institut de Recherche Servier Paris-Saclay, 3Neurology Therapeutic Area, Institut de Recherche Servier Paris-Saclay
Introduction/Objectives: Antisense oligonucleotide (ASO) is a promising therapeutic modality allowing to specifically modulate protein expression. Clinical development in brain disorders requires overcoming the inability of ASOs to cross the blood brain barrier with a direct administration into the central nervous system (CNS). In this context, highly mechanistic translational approaches are necessary to inform first in human (FIH) study design. In the work described here, we focused on S230813, an ASO developed to target alpha synuclein (a syn) mRNA in central nervous system after intrathecal (IT) administration. Preclinical PK and PKPD data generated in non human primates (NHP) and transgenic mice with this ASO were integrated in a single minimal PBPK-PD approach [1] and informed potential clinical designs while investigating the impact of some assumptions on the predictions. Methods: PKPD characterization of S230813 effect on Target Engagement (TE) in CNS was based on in vivo studies performed in a transgenic knock-in mouse model expressing human a syn gene, where a syn mRNA and protein expression was measured concomitantly to ASO concentrations. Sequential PKPD modeling approach was implemented independently for three brain structures of interest: cortex, cerebellum and basal nucleii, following an approach previously described [2]. a syn mRNA and protein were described using turnover models, with protein Rin being a function of mRNA amount, and ASO concentration increasing mRNA kout, consistent with the ASO mechanism of action. The PK profile of S230813 in CNS and periphery after IT administration was characterized using concentration measurements performed in plasma, cerebrospinal fluid (CSF) and CNS structures of interest, from a dose ranging in vivo study including 16 NHPs. Minimal PBPK (mPBPK) model described in [3] was extended to include additional compartments representing structures of interest for a syn related diseases, including basal nuclei and most transfer constants, were reestimated. Finally, the effect of the administration procedure (2mL injection volume no flush vs 1mL injection volume + flush) was tested as covariate on model parameters driving the distribution of the ASO to the CNS and the release to plasma circulation. Scaling of mPBPK model to human was performed using physiological values from literature for brain structure and biofluid volumes. Scaled mPBPK-PD model was then used to simulate a syn protein levels, considering scenarios for the different relevant brain structures and injection procedures in NHP. Model building and simulations were performed using Monolix Suite 2023R1. Results: Concentration effect relationship of S230813 on a syn mRNA was well described using considered model structure. Power concentration effect relationship was selected to describe the effect of the ASO concentration in cortex, when linear concentration effect relationship was sufficient in cerebellum and striatum, with the addition of transit compartments to account of delay. Acceptable description of protein levels assessed with VPC was obtained assuming protein Rin directly proportional to a syn mRNA expression level in considered structures. Extended mPBPK model was able to well describe the PK profile of S230813 after repeat IT administration. Concentration levels in cortex and cerebellum were similar to the ones described using the native mPBPK model. The distribution rate to brain structures was 2-fold lower with the 2mL injection volume compared to 1mL injection volume + flush, leading to an approximative same change magnitude in CNS tissue concentration in the three structures of interest. Simulations performed using the mPBPK-PD model scaled to human suggested that a dose of S230813 30mg monthly (QM) was required to reach 50% target protein inhibition in cortex, while 80mg QM was required to reach the target in cerebellum and basal nuclei. This discrepancy resulted primarily from the distribution gradient of the ASO in the different CNS structures, and secondarily from the difference in concentration-effect relationship across CNS structure estimated in transgenic mice. When simulated (assuming CNS distribution similar to the one observed with the 2mL injection volume procedure) target inhibition of a syn protein in basal nucleii and cerebellum could not be reached with doses =100mg QM. Conclusions: Modeling integration of biodistribution data allow to quantify the impact of relevant CNS structures and direct-to CNS injection procedure on human dose prediction, potentially affecting clinical design and choice of dose range to explore in FIH.
[1] Gao et al., Predicting the pharmacokinetics and pharmacodynamics of antisense oligonucleotides: an overview of various approaches and opportunities for PBPK/PD modelling Expert Opin Drug Metab Toxicol. 2023 Dec;19(12):979-990. doi: 10.1080/17425255.2023.2283524. [2] Madelain et al., Multicompartmental PKPD model to inform active exposure of nucleic acid therapeutic candidate in CNS disease, PAGE 30 (2022) Abstr 10003 [www.page-meeting.org/?abstract=10003] [3] Monine et al., A physiologically-based pharmacokinetic model to describe antisense oligonucleotide distribution after intrathecal administration J Pharmacokinet Pharmacodyn. 2021 Oct;48(5):639-654. doi: 10.1007/s10928-021-09761-0.
Reference: PAGE 33 (2025) Abstr 11558 [www.page-meeting.org/?abstract=11558]
Poster: Drug/Disease Modelling - CNS