Sylvie Retout 1, Félix Jaminion 1, Ulrich.F.O. Luhmann 2, Valérie Cosson 1, Nelson Guerreiro 1
1 Roche Pharma Research and Early Development, Pharmaceutical Sciences, Roche Innovation Center Basel (Basel, Switzerland), 2 Roche Pharma Research and Early Development,Translational Medicine Ophthalmology, Roche Innovation Center Basel (Basel, Switzerland)
Introduction: Geographic Atrophy (GA), a progressive and advanced form of age-related macular degeneration characterized by irreversible retinal damage, remains an area of high unmet medical need. The alternative pathway (AP) of the complement system has been implicated in its pathogenesis. Sefaxersen (IONIS-FB-LRx, ISIS 696844) is a ligand-conjugated antisense oligonucleotide designed to selectively degrade hepatic Factor B mRNA, resulting in reduced Factor B production and attenuated systemic AP activity. The Phase 2 Golden study (NCT03815825) (N=330) was a double-masked, randomized, placebo-controlled, multiple-dose, adaptive study in GA patients. The primary efficacy endpoint was the GA lesion area growth rate from baseline to Week 49, assessed by fundus autofluorescence imaging.
Objectives: To develop a pharmacokinetic/pharmacodynamic (PK/PD) model and leverage a GA disease progression model (DPM) [1] to assess the relationship between sefaxersen PK, systemic Factor B inhibition, AP activity (measured by Wieslab assay [AH50 WAP]) and GA lesion growth.
Methods: Data were pooled from two Phase 1 studies (ISIS 696844-CS1/-CS2) in healthy volunteers (HV) and the GA Phase 2 Golden study. Sefaxersen was administered subcutaneously at 10-40 mg in HV. In GA patients, 40, 70 or 100 mg was administered subcutaneously on Days 1, 15, 29, then every 4 weeks. The PK/PD dataset included 53 HV (41 active, 12 placebo) and 329 GA patients (223 active, 106 placebo). A semi-sequential approach was taken, developing first the population PK model, and then the PK/PD model, taking into account the influence of potential baseline covariates.
The efficacy dataset included 1766 longitudinal GA lesion area observations. A GA DPM previously developed using data from GA patients in Phase 3 trials of lampalizumab [1] was leveraged. Prediction-corrected visual predictive checks were performed to assess if the established GA DPM could (i) adequately predict the Golden study GA placebo data; (ii) support a sefaxersen drug effect estimation when using the entire Golden study GA dataset.
Results: A two-compartment model with first-order absorption and elimination processes best described the PK data. Lower clearance was estimated in GA patients compared to HV and a bodyweight effect was detected, with clearance and volume parameters increasing with weight. A clear response in plasma between drug exposure, Factor B down regulation and downstream AP inhibition was observed. Indirect response sigmoid Imax models inhibiting the production of Factor B and AH50 WAP in a sefaxersen concentration-dependent manner were employed. The developed PK and PK/PD models demonstrated good predictive performances for the central tendencies and variabilities of the observed time courses for PK, FB, and AH50 WAP. Used in a simulation mode, the models showed approximately 75-80% Factor B inhibition at the 70 mg dose, correlating to a median of approximately 40% reduction in systemic AH50 WAP activity. In addition, the PK/PD relationships were shown to reach a plateau at ~70 mg, without further inhibition expected at higher doses, which was consistent with the 100 mg data.
The GA DPM, without any treatment effect, performed similarly well in predicting GA lesion growth in patients receiving placebo and in patients across sefaxersen dose groups showing reduced systemic Factor B levels and AH50 WAP activity. This is consistent with the study finding that while sefaxersen treatment provided the expected exposure-PD response, it did not translate to a statistically significant treatment effect compared to placebo.
Conclusions: Sefaxersen achieved a robust, exposure-dependent systemic inhibition of Factor B and AP activity. The lack of a relationship between systemic PD markers and the Golden study primary endpoint indicates that an achieved ~40% systemic AP inhibition is insufficient to impact GA lesion progression, suggesting the need for a higher degree of local or systemic AP inhibition to elicit a clinical response in the eye. Although the systemic effect of sefaxersen did not translate into a clinical benefit in GA patients, a clear correlation between systemic Factor B inhibition and clinical improvement in IgA nephropathy (IgAN) patients was found [2]. The developed PK/PD models, complemented by Phase 2 data in IgAN, were therefore used to support a 70 mg dosing regimen for Phase 3 IMAgINATION study (NCT05797610) in IgAN patients [3].
References:
[1] Chanu P et al. PAGE 28 (2019) Abstr 9184 [www.page-meeting.org/?abstract=9184]
[2] Barbour S J et al.Kidney Int. 2026 Mar;109(3):592-601
[3] Retout S et al. J Am Soc Nephrol 36(10S):10.1681/ASN.202537xhdepf, October 2025
Reference: PAGE 34 (2026) Abstr 12021 [www.page-meeting.org/?abstract=12021]
Poster: Drug/Disease Modelling - Other Topics