EGFR C797S: the mutation behind osimertinib resistance

Osimertinib works because it forms a covalent bond to a single cysteine in the EGFR kinase domain: Cys797. Mutate that cysteine to serine and the warhead has nothing to hook onto. That is C797S, the on-target resistance mutation that ends third-generation EGFR therapy for a meaningful slice of patients, and the reason a fourth generation of inhibitors exists at all.

The covalent bond C797S abolishes

Osimertinib carries an acrylamide warhead that undergoes Michael addition with the thiol of Cys797, sitting at the lip of the ATP pocket. That covalent tether is what gives the drug its durable, mutant-selective inhibition. Substituting serine for cysteine removes the reactive thiol, so no covalent bond forms and the drug's residence time collapses to whatever its weak reversible affinity provides, which is not enough.

C797S shows up in roughly 10 to 20% of patients progressing on second-line osimertinib (those who had T790M first) and about 6 to 12% of patients progressing on first-line osimertinib. It was first characterized by Thress and colleagues in 2015, almost as soon as the drug entered wide use.

Why allelic context decides everything

Here is the clinically decisive subtlety: when C797S co-occurs with T790M, whether the two mutations sit on the same allele (cis) or different alleles (trans) changes what can still work.

• Trans (different alleles) — one allele carries T790M, the other C797S. A first-generation EGFR TKI can hit the C797S allele and a third-generation TKI can hit the T790M allele, so a combination can still suppress the tumor. Niederst et al. (2015) showed this experimentally.
• Cis (same allele) — T790M and C797S on one molecule. No single approved EGFR TKI, and no first-plus-third-generation combination, restores control. This is the dead end that the fourth-generation programs were built for.

This is why C797S reports increasingly come with allelic phasing from ctDNA, not just a variant call. The geometry of the two side chains in the binding pocket is the whole story, and it is exactly the kind of thing you can look at structurally before committing to a sequencing strategy.

The fourth-generation chase

Two design philosophies are competing to drug the triple mutant (19Del or L858R, plus T790M, plus C797S):

• Non-covalent ATP-competitive inhibitors — give up the covalent bond entirely and bind reversibly with high affinity. BLU-945 (Blueprint), tested in the SYMPHONY phase 1/2 trial, reports more than 450-fold selectivity for C797S/T790M mutants over wild-type EGFR, and is being run both as monotherapy and combined with osimertinib. BBT-176 (Bridge Biotherapeutics) is an orally available non-covalent TKI with nanomolar potency against the triple mutant; early SYMPHONY-era reports noted tumor shrinkage in 19Del/T790M/C797S patients.
• Allosteric inhibitors — bind a pocket outside the ATP site that opens in the C-helix-out conformation, so they are intrinsically indifferent to whether Cys797 is mutated. EAI045 was the proof of concept; JBJ-09-063 is a more drug-like successor. Because they bind a different site, allosteric agents pair naturally with ATP-competitive drugs to cover multiple resistance contingencies at once.

Most of these remain investigational. The honest status as of 2026 is that no fourth-generation EGFR inhibitor has reached approval, and the field is still working out which allelic and co-mutation contexts each chemotype actually rescues.

Try the docking yourself

The cleanest way to build intuition for C797S is to look at the bond that is no longer there. Open Studio and pick EGFR from the target catalog with C797S from the mutation chips to run molecular docking against the resistant receptor. Liganx renders the wild-type and C797S structures so you can see the serine substitution sitting where the acrylamide warhead used to anchor, and a covalent docking run will simply fail to place the bond, which is the point. For candidate non-covalent binders, compare the docked score against wild-type and mutant to read the selectivity story directly.

Liganx is molecular docking online: free, browser-based, and built for exactly this mutation-aware question. If you want to try molecular docking on EGFR resistance mutations without a local install, that is the fastest path.

Primary sources

• Thress KS, et al. Acquired EGFR C797S mutation mediates resistance to AZD9291 in non-small cell lung cancer harbouring EGFR T790M. Nat Med 21, 560-562 (2015). doi:10.1038/nm.3854
• Niederst MJ, et al. The Allelic Context of the C797S Mutation Acquired upon Treatment with Third-Generation EGFR Inhibitors Impacts Sensitivity to Subsequent Treatment Strategies. Clin Cancer Res 21, 3924-3933 (2015). doi:10.1158/1078-0432.CCR-15-0560
• Lim SM, et al. BBT-176, a Novel Fourth-Generation Tyrosine Kinase Inhibitor for Osimertinib-Resistant EGFR Mutations in Non-Small Cell Lung Cancer. Clin Cancer Res 29, 3004-3016 (2023). doi:10.1158/1078-0432.CCR-22-3901