Background: Sarcoglycanopathies are severe, early-onset autosomal recessive Limb-girdle muscular dystrophies caused by pathogenic variants in SGCA, SGCB, SGCD, and SGCG. These variants disrupt the sarcoglycan (SG) complex within the dystrophin-glycoprotein complex (DGC), leading to progressive muscle weakness. While next-generation sequencing enables rapid variant detection, interpreting variants of uncertain significance (VUS)—especially rare missense variants—remains a major diagnostic challenge in neuromuscular medicine.
Objectives: To develop a high-throughput functional platform to systematically quantify the impact of all possible single-nucleotide variants (SNVs) in SG genes, providing an independent line of evidence for variant interpretation and understanding pathogenic mechanisms.
Results: We created a saturation mutagenesis library of all SGCA SNVs and measured cell surface expression via FACS in HEK-BDG cells (expressing SGCB, D&G). We obtained functional scores for 100% of SNVs using NGS. For residues 1-316, start-loss and nonsense variants were most damaging, while missense variants showed overlapping scores, pinpointing potentially damaging substitutions. Notably, variants in the last 71 C-terminal residues behaved differently: missense showed wild-type-like scores, and nonsense variants exhibited unexpectedly high surface expression, confirmed by immunofluorescence. This suggests alternative pathogenic mechanisms, such as disrupting intracellular interactions with DGC components, rather than solely impairing localization. SGCA scores aligned with gnomAD: rare variants (AF < 1.5e-05) showed diverse scores, while common variants clustered near wild-type. Our assay (AUC=0.92) outperformed all computational predictors (e.g., MAVERICK AUC=0.85). Preliminary experiments using SGCD and SGCG mini-libraries confirmed the approach, with variant abundance for known pathogenic and benign controls aligning with expected clinical classifications.
Conclusions: Our assay demonstrates strong predictive performance and is being scaled to the sarcoglycan complex and generalizable to other membrane-associated complexes. These results provide a robust framework for integrating functional data into ACMG variant classification, ultimately improving diagnostic accuracy and identifying patients suitable for targeted therapies for LGMD and other membrane protein-associated diseases.