KIT in GIST: the four-line resistance ladder

GIST was the proof of concept that small-molecule kinase inhibition could turn a chemo-refractory sarcoma into a chronic disease. Imatinib did for KIT what it had already done for BCR-ABL. But two decades on, GIST is also the cleanest illustration of how a tumor climbs a resistance ladder one secondary mutation at a time, and why the field needed four more drugs after imatinib to keep up.

The driver: KIT, not the histology

Roughly 80% of gastrointestinal stromal tumors are driven by an activating mutation in KIT (CD117), and most of the rest by the related receptor PDGFRA. The kinase is constitutively switched on, signaling through RAS-MAPK and PI3K-AKT without ever seeing its ligand. Where the mutation sits matters enormously for which drug works.

• KIT exon 11 (juxtamembrane domain) — the most common primary mutation, ~70% of cases, and the most imatinib-sensitive. The juxtamembrane region normally clamps the kinase in an inactive state; mutating it releases the clamp.
• KIT exon 9 (extracellular) — ~10% of cases, less sensitive to standard-dose imatinib. These patients are dosed at 800 mg/day rather than 400 mg.
• PDGFRA D842V — a primary activation-loop mutation that is intrinsically imatinib-resistant from day one, because the activation loop sits where imatinib needs the kinase to be inactive.

Why imatinib eventually fails

Imatinib is a type-II inhibitor: it binds the DFG-out, inactive conformation of the kinase. That is its strength and its weakness. Tumors escape by acquiring secondary mutations that either restore the active conformation or block the drug from reaching the ATP pocket. Two structural neighborhoods dominate:

• ATP-binding pocket (exon 13/14) — V654A and T670I are the recurring gatekeeper-region mutations. They sterically interfere with imatinib binding while leaving the kinase otherwise functional.
• Activation loop (exon 17/18) — D816, D820, and N822K stabilize the active (DFG-in) conformation, which a type-II inhibitor like imatinib cannot bind at all.

Critically, these secondary mutations are polyclonal: different metastases in the same patient can carry different escape mutations, which is why a single salvage drug rarely controls every lesion.

The salvage ladder

Each later-line agent was developed to cover a slice of the secondary mutation space that the previous drug missed.

• Sunitinib (Sutent) — second line, FDA approved 2006. Potent against ATP-pocket secondary mutations (exon 13 V654A, exon 14 T670I), but weak against activation-loop mutants.
• Regorafenib (Stivarga) — third line, approved 2012. Complements sunitinib by covering activation-loop (exon 17) mutations better than the ATP-pocket ones.
• Ripretinib (Qinlock) — fourth line, approved May 2020 on the INVICTUS phase 3 trial. A switch-control inhibitor that locks the activation switch in the inactive state, giving broad coverage across both exon 13 and exon 17 secondary mutations after three prior lines.
• Avapritinib (Ayvakit) — approved January 2020, but not as a sequential line. It is a type-I inhibitor that binds the active conformation, and it is the only drug effective against PDGFRA D842V, the activation-loop mutation that defeats imatinib from the start.

The structural lesson

The whole GIST story is a conformation argument. Type-II inhibitors (imatinib) need DFG-out; activation-loop mutations force DFG-in and wipe them out. Type-I inhibitors (avapritinib) bind DFG-in but are vulnerable to ATP-pocket gatekeeper mutations. Switch-control inhibitors (ripretinib) bind a different regulatory element entirely, which is why they survive longer against a heterogeneous mutational background. When you are reasoning about which mutation defeats which drug, you are really reasoning about which protein conformation each ligand requires.

Try the docking yourself

The conformational argument is much easier to see than to read about. Open Studio and pick KIT from the target catalog, then add a secondary activation-loop mutation such as D816 from the mutation chips. Dock imatinib against the wild-type pocket and then against the mutant, and watch the type-II binding mode lose its grip as the activation loop flips conformation. Molecular docking makes the selectivity story concrete in a way a sequence table never will.

Liganx is molecular docking online: free, browser-based, and built around exactly this kind of mutation-versus-drug question. If you want to run molecular docking on KIT secondary mutants without a local install, that is the fastest path to a pose you can inspect.

Primary sources

• Heinrich MC, et al. Kinase mutations and imatinib response in patients with metastatic gastrointestinal stromal tumor. J Clin Oncol 21, 4342-4349 (2003). doi:10.1200/JCO.2003.04.190
• Blay JY, et al. Ripretinib in patients with advanced gastrointestinal stromal tumours (INVICTUS): a double-blind, randomised, placebo-controlled, phase 3 trial. Lancet Oncol 21, 923-934 (2020). doi:10.1016/S1470-2045(20)30168-6
• Heinrich MC, et al. Avapritinib in advanced PDGFRA D842V-mutant gastrointestinal stromal tumour (NAVIGATOR): a multicentre, open-label, phase 1 trial. Lancet Oncol 21, 935-946 (2020). doi:10.1016/S1470-2045(20)30269-2