Novel models for the functional characterization of SDHA germline variants to predict cancer risk.

11532 Background: SDHA mutations are the most common cause of SDH-deficient GIST. More than 80% of SDHA-mutant GIST are associated with heterozygous germline mutations. Enhanced cancer surveillance of individuals carrying a known pathogenic germline SDHA mutation has the potential to detect early-stage tumors, allowing for curative surgery. However, of the > 1000 SDHA missense variants listed in ClinVar, > 95% are VUS. We must improve our ability to interpret the significance of SDHA variants before genetic sequencing can be utilized to its fullest potential for assessing cancer risk. Methods: We generated a novel clonal SDHA knock out (SDHA KO) human cell line with complete loss of SDHA expression and enzymatic activity. We next generated a clonal cell line harboring one copy of a landing pad cassette, allowing for site-specific integration of transfected expression cassettes mediated by Bxb1 recombinase. We validated that wild-type SDHA cDNA restored SDHA expression and enzymatic activity to levels equivalent to the parental HAP1 cells. We then selected 17 known benign variants (B, LB) and 21 known cancer missense variants (P, LP). In addition, we selected six missense variants associated with primary mitochondrial disease rather than cancer. Lastly, we selected 22 missense VUS. We created a mutant cell line in our KO landing pad cell line for each SDHA variant, as well as a control nonsense variant (SDHA R31X). Cell lines were analyzed for normalized SDH activity measurements (SDHA WT activity score = 1) Results: As expected, benign variants were collectively WT-like, with a mean Activity Score of 0. 917. However, there was substantial variability amongst individual variants, with SDHA Y55H having the lowest score at 0. 215. Although known PMD variants (mean Activity Score = 0. 358) were associated with lower functionality than that of benign variants, the extent of SDH dysfunction corresponding to cancer variants (mean Activity Score = 0. 007) was substantially lower. Notably, all but one cancer variant could be described as functionally amorphic, as their Activity Scores were not significantly different from that of a control null variant, SDHA R31X, whose mean Activity Score was 0. 004. We used the Activity Scores of these clinical control variants to establish thresholds to classify "cancer-like" or "benign-like" variants, which meet ACMG/AMP criteria for PS3ₛtrong and BS₃ strong evidence according to an OddsPath calculation. We biochemically profiled 22 cell lines expressing a SDHA VUS. Using functional data and ACMG criteria, 3 variants could be reclassified as "Likely Benign", 14 could be reclassified as "Likely Pathogenic", and 5 remained a VUS (77% successfully reclassified). Conclusions: Functional characterization of SDHA germline missense variants has the potential to identify individuals who will benefit from genetic counseling and enhanced cancer screening procedures.