Functional genomics reveals the mutational and epistatic basis of asciminib resistance

Drug resistance is a constantly evolving challenge. We confront this by implementing a functional genomics approach that combines the breadth of base editing with the depth of deep mutational scanning to map the resistance landscape of the allosteric ABL inhibitor asciminib. While asciminib circumvents common resistance to active-site inhibitors like imatinib by pharmacologically restoring the natural autoinhibition of the ABL kinase, its own efficacy is threatened by an evolving set of resistance mechanisms. Our work systematically dissects this threat, revealing an impressively diverse resistance landscape that extends to the SH3, SH2, and linker domains. We uncover 279 novel asciminib resistance mutations not annotated in the clinic and provide unprecedented structure-function detail. Furthermore, we conducted the first "edit-on-edit" screen to probe for resistance epistasis, identifying a rare but potent synergistic interaction between the SH3 domain (V73A) and the P-loop (Y253H) that confers high-level resistance, highlighting a potential role for complex evolutionary pathways in ABL. Critically, using a custom FRET biosensor open asciminib unbound vs closed asciminib bound state in live cells, we establish a direct, quantitative link between the degree of cellular drug resistance and the protein's conformational openness in vivo (r=0.82). Across diverse asciminib single mutations spanning the ABL kinase and regulatory domains, as well as the V73A/Y253H combination of mutations, we observe that the measurement of a single convergent biophysical state in cells explains the cellular asciminib resistance phenotype. This suggests a surprisingly strong convergence between prior structural and enzymology work on dynamic equilibria in ABL with purified proteins and our work on the degree of resistance in living cells. These findings provide a comprehensive roadmap for interpreting clinical asciminib failure and suggest that patient monitoring should extend beyond the kinase domain.