Functional analysis of genetic variants: contribution to the diagnosis of inherited metabolic diseases

Functional analysis of variants of unknown significance is gaining more importance in the development of a sound diagnostic strategy. The abundance of sequencing methods changed the molecular genetics in the field of inborn errors of metabolism. The large amount of data opens more opportunities for diagnostics but ironically, also brings more challenges in differentiating pathogenic variants from the non-pathogenic ones. Historically, to establish a biochemical diagnosis, functional diagnostic tests (like metabolite screening and enzyme assays) are used as first tier. Subsequently, targeted Sanger sequencing is employed for genetic diagnosis. Over the years, various DNA sequencing techniques are being more and more used at earlier stages in the diagnostic work up. As most of the time these investigations do not directly result in a final diagnosis, additional approaches are needed for variant classification. A great array of computational tools are currently used for the interpretation of pathogenicity of variants. However, their predictions are often in disagreement, which makes it difficult to decide which one is the most accurate. In contrast to the computational tools, functional assays have the potential to confirm or rule out variant pathogenicity. When functional tests demonstrate the pathogenicity of novel variants, in most cases, the diagnosis is confirmed. However, if the studied variants prove to be non-pathogenic, it creates awareness that further studies are necessary. This is particularly relevant for variants identified via WES/WGS in patients with a broad clinical phenotype or with a phenotype that appears to fit with the presumed pathogenic variants/genes. There is a large choice of functional studies that can be used, from targeted to untargeted approaches. The translational research presented in this thesis, has implications in patient care and is already (partly) implemented in diagnostics in our laboratory. Our work focuses on functional characterization of missense variants in genes involved in 2-hydroxyglutaric acidurias (SLC25A1 - chapter 2 and 3, and D2HGDH - chapter 4), in GABA metabolism (ALDH5A1 - chapter 5 and 6) and cerebral creatine deficiency syndromes (SLC6A8 - chapter 7). All these disorders are rare inborn errors of metabolism, for which the pathophysiology is not completely elucidated and for which no treatment is available. Our studies extended the molecular genetics of the addressed metabolic disorders, by identification of the gene associated with the D/L-2-HGA (chapter 2) and by identification of novel variants in other deficiencies (chapter 3, 4, 5, 6 and 7). Accordingly, the functional assessment of missense variants in these genes and their clinical implications are discussed in the mentioned chapters. One of the most used functional analysis technique in VUS research, is transient transfections of recombinant genes into cell models. In our experience, optimal expression of the proteins of interest can be achieved in a relatively short time (one to three days), facilitating a rapid diagnosis, especially important when time is crucial (e.g. prenatal diagnosis). Alternatively, more advanced gene editing strategies can be used. In chapter 5, we used the Flp-In system in corroboration with CRISPR Cas9 knock out technique to generate SSADH deficient HEK293-Flp-In cell lines, in which a select number of ALDH5A1 missense variants were stably expressed. The majority of the variants addressed in this thesis resulted in impaired activity, indicating that the investigated amino acid residues are essential for proper protein function. Still, the classification of missense variants with high residual activity can remain challenging. It is important to realise that the more aspects are taken into account, the more accurate the variant classification and ultimately the patient diagnosis will be. Functional data, in collaboration with broad evaluation of clinical, biochemical, in silico and genetic data, are the desired combinatorial approach for a correct diagnosis.