Multiplexed assays reveal effects of missense variants in MSH2 and cancer predisposition

Abstract

DNA sequencing plays an increasingly central role in clinical research and diagnostics. Genome-wide association studies have established many links between genes and disease but do not reveal the effect of most of the many possible variants within each disease-related gene. Thus, while the explosion in sequencing of human genomes has revealed millions of missense variants that change protein sequences, we only understand the phenotypic and clinical consequences of a minute fraction of these. This lack of knowledge has direct consequences for clinical action. Even if a variant is discovered in a known disease-related gene, most variants have the status of “unknown significance” (VUS) simply because they have not been encountered before in the population or been studied in the laboratory. Lynch Syndrome (LS) is a cancer predisposition syndrome that increases the risk of particularly colorectal and gynecological cancers [1]. LS is generally caused by loss-of-function (LoF) variants in one of several mismatch repair (MMR) genes, including MSH2 [2]. Identification of pathogenic variants in MSH2 would be of direct clinical relevance, but many missense variants are of unknown pathogenic significance. While computational methods exist to predict pathogenicity, including methods specific for MMR genes [3], they remain imperfect and are only considered as “supporting evidence” for variant classification [4]. For this reason, a number of experimental approaches have been undertaken to assess whether a specific missense variant in MSH2 is pathogenic or not [5–7]. Some methods can provide detailed mechanistic understanding, yet they can be time consuming since each variant is handled individually and further, they are most easily applied retrospectively. Thus, most current functional assays are challenging to scale to the almost 18,000 possible single amino acid substitutions in MSH2, making it difficult to assign pathogenicity to any new clinically discovered variant. In contrast, experiments based on multiplexed assays of variant effects (MAVEs; also sometimes known as deep mutational scanning) can be used to probe the effects of thousands of variants in a single experiment [8,9]. MAVEs combine developments in high-throughput DNA synthesis, functional assays, and rapid sequencing techniques. The first step in a MAVE is to construct a DNA library of variants that can be introduced into cells, e.g., by integration on the chromosome, on a plasmid, or by genome editing. The next step is to separate variants by a property of interest. This is often achieved by applying selective pressure, such that cells carrying a functional variant from the library will have higher growth rates than those with nonfunctional variants or, alternatively, by coupling to observable phenotypes like fluorescence followed by cell sorting. The relative frequencies of the variants in the library change depending on how well they are able to perform under selective conditions and are determined by DNA sequencing of the pool of cells before and after the selection. Finally, each variant’s