6 min readLiganx team

KRAS G12C — clinical landscape, 2026

A field guide to the four approved KRAS G12C inhibitors, what's known about resistance pathways, and where the medicinal chemistry is going next.

KRAS sat in the “undruggable” column for thirty years before sotorasib cleared FDA in 2021. Five years on, four covalent G12C inhibitors are approved or filed, the first resistance mechanisms have been characterized in the clinic, and the next-generation chemistry is already moving past the cysteine warhead. Here’s where the field actually stands.

The four approved compounds

All four target the same druggable switch-II pocket that opens only in the GDP-bound state of the G12C mutant. The cysteine at position 12 is the hook every covalent inhibitor exploits — wild-type KRAS has glycine there, so the wild-type protein is intrinsically spared.

  • Sotorasib (AMG 510, Lumakras) — Amgen, FDA approved May 2021 for previously-treated NSCLC. CodeBreaK-100 reported a 37% objective response rate. Hepatotoxicity is on the label; AST/ALT monitoring is mandatory.
  • Adagrasib (MRTX849, Krazati) — Mirati/BMS, FDA approved December 2022. Slightly higher CNS penetration than sotorasib (TPSA < 90 Ų) — relevant because brain mets are common in NSCLC. KRYSTAL-1 ORR was 43%.
  • Divarasib (GDC-6036) — Genentech, filed 2024. Higher selectivity for G12C over related KRAS mutants in vitro, which may translate to a cleaner safety profile.
  • Garsorasib (D-1553) — InventisBio, approved in China 2024. Independent chemotype from the others; useful as a sequencing option after first-line resistance.

What resistance looks like

Awad et al. (2021) and Tanaka et al. (2021) characterized the first sotorasib failures and the resistance landscape is heterogeneous, not dominated by a single “gateway” mutation the way EGFR T790M was for first-gen EGFR inhibitors. The recurring motifs:

  • On-target second-site mutations — Y96D, R68S, H95D, H95Q. Each disrupts the switch-II pocket geometry that the inhibitor relies on. These are the cleanest signal that a drug binds where you think it binds.
  • KRAS amplification — straightforward gene-dosage escape. The drug is doing its job, there's just more target.
  • Bypass-track activations — co-occurring mutations in MET, BRAF, NRAS, MAP2K1. Pathway pivots that route around the inhibited node entirely.

The clinical implication: G12C inhibitor monotherapy has a ceiling. Combinations with SHP2, SOS1, MEK, or anti-PD-1 are most of the ongoing trials. The chemistry implication: the next wave is going after the GTP-bound state (the “ON” conformation), not just the GDP-bound state these four work on.

What's coming next

Three threads are worth watching. Pan-KRAS inhibitors(e.g. RMC-6236) bind the GDP-bound state non-covalently and hit G12D, G12V, and G13D in addition to G12C — which would address the 80% of KRAS-mutant tumors the covalent G12C drugs miss.GTP-state binders (RMC-7977, BI-2865) target the ON conformation directly and short-circuit the GTP-loading step rather than waiting for the cycle to land in GDP. And protein-degraders (KRAS-targeting PROTACs) are in early development at Arvinas and Foghorn — degrading the protein avoids the resistance-via-second-site-mutation pattern entirely, since you can't second-site-mutate something that's not there.

Try the docking yourself

The canonical KRAS G12C structure for sotorasib is 6OIM — switch-II pocket open, sotorasib covalently bound to Cys12. Open Studio and pick KRAS from the target catalog with G12C from the mutation chips to dock your own ligands against the same structure. Liganx renders both the wild-type and G12C receptors side-by-side so you can see the selectivity story directly — most G12C-selective compounds will score 1-2 kcal/mol better against the mutant. The ADMET panel will flag hepatotoxicity (the sotorasib pattern) if your candidate has the same liabilities.

Liganx is molecular docking online: free, browser-based, and set up for exactly this kind of mutation question. If you want to try molecular docking on KRAS G12C without a local install, that is the fastest path.

Primary sources

  • Canon J, et al. The clinical KRAS(G12C) inhibitor AMG 510 drives anti-tumour immunity. Nature 575, 217–223 (2019). doi:10.1038/s41586-019-1694-1
  • Awad MM, et al. Acquired Resistance to KRAS G12C Inhibition in Cancer. NEJM 384, 2382–2393 (2021). doi:10.1056/NEJMoa2105281
  • Tanaka N, et al. Clinical acquired resistance to KRASG12C inhibition through a novel KRAS switch-II pocket mutation and polyclonal alterations converging on RAS-MAPK reactivation. Cancer Discov 11, 1913–1922 (2021).