PREreview del Anticipating on-target resistance to WRN inhibitors in microsatellite unstable cancers

Orcholski et al. present a thorough exploration of acquired resistance mechanisms to a covalent and a non-covalent inhibitor of WRN helicase. The study provides a comprehensive overview of resistance hotspots within WRN and identifies genes that potentiate WRN inhibitor sensitivity, offering insights into co-treatment strategies and rational drug selection as resistance emerges. The synthetic lethality of WRN in MSI cancers stems from a hypermutator phenotype, making it a highly relevant target for understanding resistance mechanisms as clinical use advances. The combination of base editing, which can introduce mutations at very distant endogenous sites, with full-coverage deep mutational scanning of a specific region via an introduced library provides a valuable resource for anticipating the specific resistance mechanisms likely to be encountered clinically.

Minor Points:

1. The inability to recover homozygous mutant clones is an interesting outcome and suggests that the resistance mutations impair WRN function. Is it possible to generate a compound heterozygous clone carrying two different resistance mutations that act through distinct mechanisms? This could provide insight into how two partially deficient WRN alleles cooperate.

2. The DMS screen reveals informative mutational effects, but additional analyses in the supplementary material would strengthen the mechanistic interpretation beyond active-site contacts.

   1. Within the binding site (residues 727–737), are there patterns in the types of substitutions that confer resistance? It appears that charged residues and serine/threonine substitutions have larger effects. Evaluating the residue-type sensitivity between the two inhibitors would be informative.

   2. A difference map showing the log fold change between the two inhibitors would provide an additional view of drug-specific versus general loss-of-function mutations and may offer more insight into the distinct mechanisms of action for the inhibitors.

   3. Not all mutations at C727 confer resistance in the VVD-214 DMS screen. Given that VVD-214 is a covalent inhibitor targeting this residue, why do substitutions such as serine and alanine at C727 not produce full resistance?

3. In Figure 3F. The effect on WRN levels of the C727V mutation looks to be more pronounced for the HRO761-treated sample than the VDD-214-treated sample. Since VVD-214 is a covalent inhibitor that binds at C727, the effect on WRN detected appears to be opposite for this mutation. How is that so?

   1. The differences in structural views in figure 3 also make it

4. The differential resistance mechanisms between the two inhibitors could be more clearly illustrated by mapping mutations onto the protein structure, grouped by whether they affect one or both inhibitors. The four mutations in Figure 3 would work well as a summary structural panel.

Editorial points

1. In Figures 1B–C, the dot plot shows all tested mutations but does not convey library coverage. A heatmap representation of the base editing library, analogous to the DMS format, would allow more direct comparison between the two approaches and make differences in coverage immediately apparent.

2. In Figure S1G, residue 731 is labeled as D731=, but the wild-type sequence shown above the panel indicates E731.

3. The citation for the DrugZ method is missing from the reference list.

Another preprint (https://www.biorxiv.org/content/10.64898/2025.12.23.695783v1.full ) has some similar analyses and both strongly implicated SMARCAL1. But they differ in the specifics of DNA damage response/NEHJ pathway genes implicated in resistance with implications for companion therapies. They also differe in the discussion of the importance of P53. The authors may wish to comment about these differences and weather they result from different cell backgrounds, analyses, or other factors.

Competing interests

The authors declare that they have no competing interests.

Use of Artificial Intelligence (AI)

The authors declare that they did not use generative AI to come up with new ideas for their review.