Beyond the promise and potential of base editing in modelling or correcting known pathogenic mutations, two comprehensive and recently published papers have demonstrated the power of mutational tiling of endogenous loci using base editors. These studies have enabled the identification of mutations of functional importance as well as the discovery of drug-variant interactions that will ultimately guide and accelerate the path towards drug discovery and precision medicine. Combining base editing with functional screening will help uncover novel therapeutic targets and elucidate genetic underpinnings of disease or drug response, ultimately advancing the path towards precision medicine.
While saturation mutagenesis screens and saturation genome editing (SGE) have been used to inform on molecular determinants of human disease, these strategies remain technically challenging and limited in their scope and application. These limitations can be overcome using CRISPR-derived base editors, which require a single guide RNA to target loci of interest and introduce precise C-to-T (cytosine base editors, CBE) or A-to-G (adenosine base editors, ABE) transitions.
Two research groups led by John Doench1 and Alberto Ciccia2 recently showcased the value of base editing to analyze the impact of single nucleotide variants (SNVs). Hanna, Doench and colleagues initially benchmarked the performance of CBEs in positive and negative screens and successfully identified loss-of-function (LOF) and gain-of-function (GOF) mutations in BRCA1/2 across multiple CBE-expressing cell lines. Cuella-Martin, Ciccia and colleagues corroborated these findings through identification of variants of pathogenic/likely pathogenic effect localized to critical protein domains of the same genes, indicating the robust and reproducible nature of this base editing pooled screening approach.
Novel point mutations come to light
Hanna et al. (2021) subsequently probed drug-protein function interactions and identified novel SNVs in MCL1, BCL2L1 and PARP1 that confer LOF and resistance to multiple inhibitors which target the proteins encoded by these genes. Negative screens revealed point mutations that sensitize cells to drug treatment, variants that are difficult to detect with CRISPR-Cas9 knockout screens that work on the premise of gene disruption and protein loss. Further analyses indicated that a sgRNA targeting BCL2L1 generated the intended editing profile at the target site along with two occurrences of detectable C-to-T editing within the proximal upstream region of the protospacer. This stresses the importance of hit validation from primary screens and not relying on predictive heuristics alone. Unexpected edits might also be identified when sequencing the locus targeted by the sgRNA. This finding was unusual among the sgRNA that were validated from the screen but highlighted likely locus-specific patterns and the importance of validating causal effects. Low frequency editing outside of the expected window will have little impact in negative selection screens but may carry much more weight in strong positive selection screens, underscoring the importance of determining and validating the actual edit that confers the function or phenotype of interest.
Characterization of genotype-to-phenotype associations
Both groups applied their base editing screens to specifically characterize the functionality of DNA Damage Repair (DDR) gene variants. Cuella-Martin et al. (2021) used BE3-expressing human mammary epithelial cells and a library of ~37,000 sgRNA for mutational tiling of a network of 86 DDR genes with subsequent analysis of variant effect under selective pressure of four different DNA damaging agents. The authors reported identification of LOF, GOF and separation of function (SOF) mutations in 53BP1, including SOF mutations in the tandem Tudor domain (TTD) of 53BP1 that uniquely abolish (V1540I, G1560K) or enhance (G1593K) 53BP1-dependent p53 regulation without interfering with the protein’s function in DNA double strand break (DSB) repair. A subset of point mutations in the TRAIP ubiquitin ligase enabled characterization of a novel but evolutionarily conserved TRAIP domain whose loss promotes resistance to the topisomerase I (TOPI) inhibitor camptothecin. Cuella-Martin and colleagues also identified variants in the CHEK2 kinase encoding gene that have pathogenic-like behavior.
The vast majority of SNVs in the ClinVar database lack functional characterization and warrant large-scale screening strategies to elucidate genotype-to-phenotype associations. Hanna et al. (2021) designed 68,526 sgRNA targeting 52,034 clinically relevant gene variants and screened this library in in two independent cancer cell lines. Parallel low dose treatment with two genotoxic agents allowed identification of novel LOF variants required for DNA damage repair or resistance to translational stress. Importantly, both groups used CBEs so the editing precision and range of accessible target nucleotides were limited to the specific activity of these CBEs. Of the 388,496 SNVs of unknown functional significance that Hanna et al. (2021) found in the ClinVar database, a library of sgRNA could only be designed to model 13% of these.
More potential and opportunity with more base editors to choose from
The use of alternative Cas enzymes will extend the repertoire of targetable SNVs through differing PAM dependencies, and the use of more precise deaminase effectors will improve product purity and limit the functional outcomes attributable to bystander editing within or outside of the editing window. ABEs that introduce A-to-G transitions will further broaden the range of targetable SNVs within the ClinVar database, and the recent reports of C-to-G base editors and dual base editors that catalyze both C-to-T and A-to-G mutations may also increase editing opportunities. It is anticipated that the vast array of base editors as well as innovative combinatorial strategies within pooled screening approaches will offer many opportunities for interrogating specific gene networks or variants on a whole genome level.
Together, these studies demonstrate that large-scale pooled BE screens can separate pathogenic from benign mutations and identify variants that are associated with altered cellular growth or response to DNA-damaging agents. Large-scale pooled screening coupled with base editing offers a high degree of genotype–phenotype resolution and the fast-paced development and evolution of both technologies will extend the breadth and profiling of human genetic variants and inform rational design of future drugs.
Written by Jennifer Harbottle
Jennifer Harbottle holds a PhD in cell biology from the University of Aberdeen.
Learn about Horizon's Pin-point™ base editing platform.
For more resources on base editing see these articles:
- Cell therapy engages stealth mode to target solid tumors
Therapeutics get on point with base editors - blog article
- Hanna, R. E., Hegde, M., Fagre, C. R., DeWeirdt, P. C., Sangree, A. K., Szegletes, Z., Griffith, A., Feeley, M. N., Sanson, K. R., Baidi, Y., Koblan, L. W., Liu, D. R., Neal, J. T., & Doench, J. G. (2021). Massively parallel assessment of human variants with base editor screens. Cell, 184(4), 1064-1080.e20. https://doi.org/10.1016/j.cell.2021.01.012
- Cuella-Martin, R., Hayward, S. B., Fan, X., Chen, X., Huang, J.-W., Taglialatela, A., Leuzzi, G., Zhao, J., Rabadan, R., Lu, C., Shen, Y., & Ciccia, A. (2021). Functional interrogation of DNA damage response variants with base editing screens. Cell, 184(4), 1081-1097.e19. https://doi.org/10.1016/j.cell.2021.01.041