Mutational spectrum of the SLC12A3 gene in a cohort of patients with Gitelman syndrome in north-eastern Italy

Introduction: Gitelman syndrome (GS) is an autosomal recessive renal tubulopathy characterized by hypokalemic metabolic alkalosis, hypomagnesemia, and hypocalciuria. Biallelic loss-of-function variants in the SLC12A3 gene, encoding the thiazide-sensitive Na⁺-Cl⁻ cotransporter (NCC), underlie the biochemical signature of GS and its clinical variability: genetic testing is the definitive tool to establish the diagnosis and to differentiate the condition from overlapping phenotypes. Structural studies have provided a molecular framework to interpret the numerous pathogenic variants which have been reported, but genotype-phenotype correlations remain elusive. Objectives: This study aimed to characterize the spectrum of SLC12A3 variants in a cohort of Italian patients with a clinical diagnosis of GS, to compare local allele frequencies with public databases, and to investigate novel variants. Materials and methods: The analyses were performed using a custom Agilent SureSelect NGS panel including SLC12A3 and other genes to be considered in the the differential diagnosis. After sequencing on Illumina platforms, we employed bioinformatics tools for variant calling and clinical interpretation (SureCall, Dragen, and Emedgene), pathogenicity prediction and classification (Franklin and Varsome), splice site prediction (Alamut, Genscan, and HSF), cross-species alignments (BLASTp), and molecular modeling (PyMOL). CNVs were validated by MLPA. Results and discussion: The entire SLC12A3 coding region was sequenced at a depth of at least 20× in 49 patients (25 males and 24 females, aged 8-84 years). The analysis was conclusive in 40 cases (82%), 39 carrying biallelic SLC12A3 variants; a single variant was found in 5 cases and none were detected in 4. One patient harbored biallelic pathogenic SLC12A1 variants, and one of the singlevariant cases also carried a pathogenic CLCNKB variant. Among the SLC12A3 patients, 32 were compound heterozygotes and 7 homozygotes. Overall, 51 alleles were identified, 37 known and 14 novel. They comprised 34 missense and 3 nonsense changes, 5 affected splicing consensuses, 7 were frameshifts, and 2 caused multi-exon deletions. The most frequent variant was p.(Cys994Tyr); variants common in the European population, such as p.(Gly741Arg) and p.(Ala13Pro), were rare in this cohort. The 14 novel variants included 9 missense, 2 frameshifts, 2 splice site alterations, and 1 multi-exon deletion. Bioinformatics tools, conservation analysis, and 3D modeling supported the pathogenicity of the novel missense changes. The splice site variants were classified as pathogenic by multiple prediction tools; one was also validated through a functional assay. The pathogenicity of the frameshift variants and of the multi-exon deletion is obvious. Patients with a single variant may harbor non-coding region alterations or undetectable CNVs, and those without any variants could be affected by nongenetic conditions. Conclusions: This study showed that the mutational landscape of Italian GS patients partly differs from the general European population. Several novel variants were identified and classified as pathogenic/likely pathogenic, and molecular modeling provided additional support for their interpretation. The NGS panel proved effective for GS characterization. Future improvements should target the promoter, deep intronic variants, and the breakpoints of other gross rearrangements.