Phase 3 FACETS study of migalastat HCl for Fabry disease: post hoc GLA mutation-based identification of subjects likely to show a drug effect

Abstract

Fabry disease is caused by more than 800 different mutations in the GLA gene that encodes the lysosomal enzyme α-galactosidase A (α-Gal A), resulting in accumulation of globotriaosylceramide (GL-3) and end-organ damage. Migalastat HCl (1-deoxygalactonojirimycin HCl, AT1001, GR181413A) is an investigational pharmacological chaperone that selectively binds and stabilizes α-Gal A, leading to increased cellular levels and greater lysosomal activity. Previously, a “research” human embryonic kidney (HEK) cell-based in vitro assay was developed to identify migalastat-responsive mutant forms of αGal A. The FACETS study (NCT00925301) compared migalastat to placebo in subjects with GLA mutations categorized as “amenable” to migalastat based on the research in vitro assay available at study initiation. The primary endpoint of this study was the percent change from baseline in kidney interstitial capillary (IC) GL-3 inclusions (responder analysis, 50% reduction threshold), which was not met. Meanwhile, a GLP-validated assay was created (“GLP HEK assay”) and testing of all mutant forms was completed before unblinding of the 6-month data. Approximately 90% of mutations remained in the same category of “amenable” or “non-amenable” using the same predefined in vitro criteria. However, 15 of 67 subjects enrolled in the study express mutant forms of α-Gal A considered to be nonamenable based on the GLP HEK assay, including 8 subjects with mutation R342Q. These 15 subjects may not be sufficiently responsive to migalastat, and may have impacted the previously reported study results. To explore the effect of migalastat in subjects with GLP HEK amenable mutations only, a post hoc analysis of the change from baseline (difference) in kidney IC GL-3 inclusions was conducted using an ANCOVA model with covariate adjustment for the baseline value and treatment-by-baseline interaction. Analysis of the quantitative difference and inclusion of baseline values in the analysis is the most appropriate measure of an effect on kidney IC GL-3 as many values were low at baseline. This analysis was first applied to all subjects in themITT population (n = 30 per group). Themean change from baseline in the migalastat group was a decrease of 0.22 ± 0.11 inclusions per capillary (IPC) (-66 inclusions per 300), and the placebo group showed an increase of 0.06 ± 0.09 IPC (+18 per 300); p = 0.052. The data were also analyzed for subjects with GLP HEK amenable mutations only (n = 25 migalastat, n = 21 placebo). The change from baseline in the migalastat group was a decrease of 0.31 ± 0.12 IPC (-93 per 300), while the placebo group showed an increase of 0.10 ± 0.13 IPC (+30 per 300); p = 0.0024. These findings were observed in the presence of a significant treatmentby-baseline interaction. This analysis demonstrates a measurable drug effect on change from baseline in kidney IC GL-3 inclusions during the first 6 months of treatment with migalastat, which is more pronounced in subjects with GLP HEK amenable mutations and higher baseline values. Additional data will be presented to support use of the GLP HEK assay to identify Fabry disease subjects with mutations that are most likely to demonstrate a drug effect with migalastat HCl.