Introduction:
Patients with neurodevelopmental disorders (NDDs) struggle to access medical care and genetic counselling as it can be difficult to pinpoint the genetic cause of these disorders. Mutations in the SWI/SNF chromatin remodeling complex are among the most frequent causes of NDDs. Genetic databases such as ClinVar, Simons Foundation Powering Autism Research (SPARK), and the Simons Searchlight Collection (SSC) report over 2500 mutations that affect one of the 29 SWI/SNF subunit genes. The American College of Medical Genetics has published guidelines for classifying variants as pathogenic or benign. Functional testing may show that a variant has a deleterious effect on the protein, providing strong supporting evidence for a pathogenic classification. Given that the SWI/SNF complex is the most frequently mutated complex in NDDs, it is important to develop efficient functional assays to screen variants of unknown significance (VUS) in SWI/SNF subunits. In this study, a novel Drosophila-based assay will be used to determine the functional effect of VUSs in the subunit gene SMARCD1. The study aims to provide functional data that can establish their classification as pathogenic or benign.

Methods:
Bap60, the Drosophila orthologue of SMARCD1, is necessary for fly survival. CRISPR has been previously used to insert a KozakGal4 cassette into the Bap60 locus on the X chromosome. Males or homozygous females with this insertion are not viable due to the disruption of the Bap60 coding sequence. Additionally, the KozakGal4 cassette allows for the expression of Gal4 under the control of Bap60 regulatory elements. In the humanised rescue assay, heterozygous Bap60-KozakGal4 females are crossed to UAS-SMARCD1 males. The resulting progeny express GAL4, which can bind to the UAS to drive expression of SMARCD1. We showed that the expression of SMARCD1 rescues lethality of Bap60 null mutants due to functional conservation between the two proteins. Variants of SMARCD1 will not rescue lethality if they have a detrimental effect on protein function, but they will rescue lethality if the function is mainly unaffected. Consequently, the readout for this assay is percent survival. This study will test 25 VUSs that have been identified in databases such as ClinVar, SSC, and SPARK.

Results:
Following the optimisation of assay conditions, we found that the expression of human SMARCD1 rescues survival of Bap60 null mutants from 0% to 76.4% ± 7.45% (mean ± standard error of the mean). The assay was calibrated with two truncated versions of SMARCD1 which serve as positive controls: a deletion of amino acids 1-290 and a deletion of the SWIB domain, which is responsible for binding to other subunits of the SWI/SNF complex. As expected, both deletions resulted in 0% survival. Additionally, a known pathogenic variant that causes Coffin-Siris Syndrome, c.1457G>A (p.Trp486*), was shown to have 0% survival. These results suggest that the humanized rescue assay is effective at detecting loss of function variants of SMARCD1. To date, we have tested nine of the 25 identified VUSs. Four of these were found to have significantly decreased survival compared to the reference: c.699C>A (p.Phe233Leu, 2.56% ± 2.22%, p=0.0015), c.907C>T (p.Arg303*, 0% ± 0%, p<0.0001), c.1112T>C (p.Ile371Thr, 20.93% ± 5.09%, p=0.0093), and c.1343del (p.Arg448Leufs*19, 0% ± 0%, p<0.0001). Testing of additional SMARCD1 variants is underway.



Conclusion:
This study presents a robust Drosophila-based humanised rescue assay to assess the functional effect of SMARCD1 VUSs. Thus far, four variants were shown to decrease survival compared to the reference protein. This provides strong supporting evidence for the classification of these VUSs as pathogenic. The assay can be used to test additional NDD-implicated variants as they are identified and may be adapted for other SWI/SNF subunits that are conserved between humans and Drosophila, such as SMARCA2.