Jannik Vollmer 1, Charlotte Bon 2, Till Mesmer 2, Matthias Machacek 1, Hans Peter Grimm 2
1 LYO-X AG (Basel, Switzerland), 2 Roche Pharmaceutical Research and Early Development, Pharmaceutical Sciences, Roche Innovation Center (Basel, Switzerland)
Objectives
Antisense oligonucleotides (ASOs) have emerged as a promising therapeutic modality for disorders of the central nervous system (CNS) [1]. Intrathecal (IT) administration is recognized as an important approach for delivering ASOs to the brain as only this route of administration effectively bypasses the blood-brain barrier (BBB) which excludes most systemically administered oligonucleotides from the CNS [2]. However, even after IT administration there is substantial inter-individual variability (IIV) in brain exposure [3] while the mechanisms underlying this variability are only poorly understood. Thus, achieving adequate and consistent brain delivery therefore remains a major challenge.
In this study, we integrated spatio-temporal 89Zr-PET imaging data with anatomy-level pharmacokinetic (PK) modeling to characterize the biodistribution of rugonersen following IT administration in healthy volunteers. The ambition was to infer the dominant processes of the distribution kinetics of the ASO, to do so with anatomical resolution, and further to pinpoint the sources of IIV. Technically, the objective was to find a model satisfying pharmacometric criteria while prescribing the anatomically and physiologically plausible connectedness of its constituent compartments. The resulting model is expected to guide the development of optimized intrathecal delivery for a broad class of ASOs.
Methods
89Zr labeled rugoenerson was administered to 24 healthy male volunteers across four cohorts testing different intrathecal administration procedures. PET imaging was performed at approximately 1, 16, 88, 168, and 288 hours post-dose. Regions of interest (ROI) were defined
for brain compartments (cerebrospinal fluid, grey matter, white matter) and vertebral levels C3 to L5. Radioactivity in each ROI was quantified as percent injected dose (%ID).
Biodistribution data were analyzed using an anatomy-level PK model within a nonlinear mixed-effects modeling framework. The model comprised distinct compartments for each vertebral level (L5 to C3), the brain, and plasma. Within each CNS compartment, the model consisted of two sub-compartments describing ASO in the CSF (i.e. a mobile phase) and bound to tissue (i.e. a static phase). Exchange between the compartments and the fluid (CSF) and static (tissue) sub-compartments was modelled with first-order rate constants that were estimated separately for each compartment.
Results
Vertebrae-level data at 1, 16, and 88 h post dosing demonstrated rapid distribution throughout the spine, with relatively minor changes thereafter. By 1 h post dosing, rugonersen had already reached the cervical region, and the time of maximum exposure (Tmax) was 1 h for all vertebral levels. At this early time point, little difference was observed between subjects with high and low brain exposure; however, pronounced differences emerged at 16 and 88 h.
Multiple alternative model structures were evaluated, including different CSF-to-plasma outflow pathways, uni- and bidirectional CSF transport, reversible or irreversible tissue binding, and spatial varying or non-varying transport rates between the (sub-)compartments. The best-performing model described the data with unidirectional CSF transport toward the brain with spatially varying rates, irreversible tissue binding at each vertebra, reversible tissue binding in the brain, and CSF-to-plasma transport occurring exclusively from the brain. Further, elimination from vertebral tissue compartments was required to adequately describe the data.
Parameter estimates revealed systematic differences between high- and low-brain-exposure groups, leading to inclusion of brain exposure group as a covariate on CSF tissue elimination, brain-to-plasma transport, and plasma clearance. The model provided a good description of the data at both the population and individual levels. Notably, independently estimated transport rates between vertebral compartments exhibited a clear spatial gradient, with slow transport in the lumbar region and progressively faster transport toward cervical levels.
Conclusions
By integrating spatio-temporal PET imaging data with anatomy-level PK modeling, we identified key processes governing ASO distribution and sources of inter-individual variability in brain exposure. The model proposes that transport from the CNS to plasma is primarily occurring from the brain and that there is an increasing speed of transport in the cervical region. We expect these findings to be not specific to this particular molecule but to be applicable to the wider class of ASOs.
References:
[1] McDowall S, et al. Antisense oligonucleotides and their applications in rare neurological diseases. Front Neurosci. 2024;18:1414658
[2] Barker SJ, et al. Targeting the transferrin receptor to transport antisense oligonucleotides across the mammalian blood-brain barrier. Sci Transl Med. 2024 Aug 14;16(760):eadi2245. doi: 10.1126/scitranslmed.adi2245. Epub 2024 Aug 14. PMID: 39141703.
[3] Łusakowska A, et al. Long-term nusinersen treatment across SMA severity: real-world experience. Orphanet J Rare Dis. 2023;18:230.
Reference: PAGE 34 (2026) Abstr 12040 [www.page-meeting.org/?abstract=12040]
Poster: Oral: Other Topics