A comprehensive framework for the interpretation of TTN missense variants

BACKGROUND: Missense variants in TTN pose a major challenge in genetic diagnostics due to their high frequency in the general population, the large size of the gene, and the complex multidomain architecture of the titin protein. While the contribution of truncating variants (TTNtv) to titinopathies is well established, the role of rare TTN missense variants remains poorly defined. Advances in computational prediction and functional testing offer new tools to assess their potential pathogenicity, which however are currently not fully utilized for clinical application. METHODS: We analyzed an international cohort of unsolved myopathy cases selected based on the presence of a rare missense variant in trans with a TTNtv. Clinical data were collected from neuromuscular centers worldwide. In silico predictions were generated using AlphaMissense and complemented by minor allele frequency (MAF) and exon usage information. Additional inclusion criteria were based on a MAF < 0.010 and an AlphaMissense score ≥ 0.792 for the missense variants, in accordance with the latest ClinGen guidelines. Selected missense variants were characterized in vitro through protein expression and cell imaging assays to assess their effects on domain solubility and aggregation. RESULTS: Thirty patients with TTNtv/missense combinations were identified, presenting with heterogeneous myopathic phenotypes, ranging from congenital to adult onset. An in-depth analysis on AlphaMissense predictions highlighted those changes most frequently predicted as possibly pathogenic. Functional assays showed that three selected variants with changes to proline, located in β-sheets of Ig domains, led to impaired folding, cytoplasmic aggregation and co-localisation with proteostasis markers. In our cohort, all non-proline mutations occurred at buried sites, while some proline substitutions affected exposed residues. Notably, the variant p.(Gln7023Pro) was identified in 5 unrelated families sharing a conserved haplotype, indicating a common ancestor. This variant and the previously reported p.(Arg25480Pro) now meet ACMG criteria for classification as likely pathogenic. CONCLUSIONS: By integrating clinical, computational, and functional evidence, we propose a framework for interpreting TTN missense variants. Combining multiple lines of evidence is essential for variants’ classification and interpretation, especially given TTN complexity. Advancing diagnostic accuracy will require tailored interpretation guidelines and a global effort in data sharing and functional validation.