A functional comparison of readthrough agent ELX-02 across a wide range of nonsense CFTR variants

BackgroundNonsense variants in CFTR account for ~10% of cystic fibrosis (CF) variants and cannot be treated with approved CFTR modulators. Translational readthrough agents such as ELX-02 offer a potential therapeutic strategy, but clinical trials evaluated mainly in G542X CFTR nonsense variant and underlined limited efficacy. This study aimed to evaluate ELX-02-mediated CFTR rescue across a broad range of nonsense variants using patient-derived intestinal organoids (PDIOs) to define variant-specific determinants of readthrough efficacy and assess its potential across a genetically diverse CF population. MethodThe ex vivo response to ELX-02 was assessed in 206 PDIOs carrying heterogeneous nonsense variants. CFTR function was quantified using forskolin-induced swelling (FIS) assay after 48-hour exposure to ELX-02. Responses were analysed by genotype and stop codon identity, with secondary validation performed in a selected subset of PDIOs (n = 60). ResultsELX-02-mediated CFTR rescue varied markedly, ranging from responses approaching those observed with approved CFTR modulators (LUM/IVA) to responses at or below detection limit. Overall, maximal responses were modest and at the lower end of the functional range for CFTR modulators. Rescue was dose-dependent and higher in PDIOs carrying two nonsense variants when compared with PDIOs carrying a single nonsense variant combined with a residual or minimal function variant. Nonsense variants in nucleotide-binding domain 1, including G542X, S466X, G550X and R553X, showed relatively higher responsiveness. ConclusionELX-02 induces limited and highly heterogeneous CFTR rescue across nonsense variants. PDIO-based functional screening provides a framework to guide patient selection and stratification for future readthrough therapy trials.