Predicted Succinated Dehydrogenase Subunit Variant Pathogenicity: Why Are SDHB Variants “Bad”?

Abstract

Abstract Variants in the 4 genes encoding subunits A-D of succinate dehydrogenase (SDH) are associated with paraganglioma and pheochromocytoma. Intuitively, loss-of-function variants affecting any of the subunits should equally diminish SDH function leading to succinate accumulation and tumorigenesis after loss of heterozygosity. However, variants in SDHB are associated with a higher prevalence of metastatic disease and a more aggressive clinical course. Evaluation of the SDH protein structure shows the fraction of amino acids in contact with other subunits or essential prosthetic groups to be: 13% (SDHA), 40% (SDHB), 28% (SDHC), and 28% (SDHD). We therefore hypothesized that SDHB missense variants are more penetrant because a larger fraction alter sensitive interfaces with other SDH subunits or essential molecular features (e.g. the three SDHB iron-sulfur clusters). We also wondered if truncating variants are more common for SDHB than other subunits. To test these hypotheses, we combined three databases (Genome Aggregation Database, ClinVar-NCBI-NIH, and Leiden Open Variant Database) and our institution’s data to create a pool of all known SDH variants. We categorized variants as truncating or missense and evaluated missense variants in the context of the SDH protein structure, scoring each variant in relation to important structures/interfaces and the severity of the amino acid change. This provided an ad hoc impact score for each variant, where a higher score predicts a more deleterious effect. We compared these scores to those obtained using the “Sorting Intolerant from Tolerant” (SIFT) tool that predicts impacts of amino acid changes based on evolutionary sequence conservation. SIFT scores of 0 to 0.05 predict deleterious effects. Both mean impact and SIFT scores could be weighted for the prevalence of each variant in the population. Our database included 2333 total SDH variants: SDHA (838, 36%), SDHB (703, 30%), SDHC (381, 16%), and SDHD (412, 18%). The fractions of truncating variants were 38%, 50%, 51%, and 53% for A-D subunits, respectively. When weighted for prevalence, these fractions were 0.39%, 6.8%, 8.2%, and 0.2%. The number of truncating variants per coding region length and the distribution of locations were similar between subunits. Ad hoc impact scores for A-D subunits were 3.08, 14.9, 9.93, and 11.0, respectively and, when weighted for prevalence, were 0.28, 3.25, 6.32, and 1.15. Mean SIFT scores for subunits A-D were: 0.185, 0.162, 0.238, and 0.410 respectively, and, when weighted for prevalence, were 0.58, 0.70, 0.22, and 0.018. Our results do not support the hypothesis SDHB variants predict a worse clinical outcome because average SDHB variants are, by chance, more biochemically severe. This suggests that SDHB loss may uniquely impact SDH biochemical function.