Endocrine Abstracts

Next Generation Sequencing (NGS) is becoming widely used for genetic diagnosis. While its capacity for detection of human genetic variations (GV) is outstanding, drawbacks is the identification of numerous GV of which functional significance cannot be predicted in silico by computer programs (variants of uncertain significance, VUS). Currently, only sensitive (but also expensive and time-demanding) in vitro cell assays, can trustfully ascertain pathogenicity of VUSs, generally of selected proteins o signalling pathways, seriously hampering efficiency of genetic diagnostic for the clinic.

Objective: To apply 3D mapping and in silico predictions as tools to inform pathogenicity of GVs identified in hypothyroid patients with alleged iodotyrosine deiodinase deficiency, as compared with in vitro dehalogenation assays.

Methods: Nine missense DEHAL1 variants identified in goitrous hypothyroid patients were subjected to both in vitro functional testing and in silico 3D modelling and docking with substrates (MIT, DIT) and cofactor (FMN), using available X-Ray crystallographic information of dimeric DEHAL1. Three (K258N, V265M and R279S) were identified in patients with undisputed diagnosis of iodotyrosine deiodinase deficiency through in vivo ¹²³I-MITdeiodination test. The rest (N108S, A202T, R246Q, L260P and E271K) are harboured by patients with clinical suspicion of DEHAL1 defect. The functional assay involved mutagenesis of DEHAL1 variants in expression vectors, transfection in HEK293 cells, addition of MIT, FMN and NADPH to culture medium and determination of % MIT decrease by LC/MS-MS.

Results: The functional assay showed significant decrease of deiodination for K258N (60%) and R279S (58%) versus WT. However, V265M did not, suggesting limitations of our assay to detect less severe mutations. The rest of variants showed normal deiodination with the exception of L260P, which behaved as deleterious (10%). Interestingly, in silico studies on K258N, R279S, V265M and L260P revealed that the amino acids changes had damaging effects on the structural stability, electrostatic properties or interaction with cofactor and substrates (FMN and MIT) on the DEHAL1, and provide a structural explanation for the iodotyrosine deiodinase deficiency. Besides, the peripheral location of N108S, A202T, R246Q, and E271K variants in the model correlates with the benign nature of such changes in the assay.

Conclusions: Structural modelling and molecular dynamics is a valuable tool to discriminate pathogenic vs VUS changes in human DEHAL1, showing consistency with the in vitro assay and superior sensitivity to detect some pathogenic changes. An additional advantage of this filtering approach is the prediction of the intrinsic molecular mechanism driving the functional damage of mutations.