BCR-ABL in CML: from imatinib to the allosteric STAMP inhibitor

Chronic myeloid leukemia is the disease where targeted therapy was born. Imatinib turned a uniformly fatal leukemia into a condition many patients die with rather than from. But the twenty-five years since have been a slow structural argument about a single kinase pocket, one famous mutation that defeats almost everything aimed at that pocket, and a final drug that won by refusing to fight there at all. If you want a clean case study in how protein structure dictates drug sequencing, BCR-ABL is the one to study.

The driver: a fusion that never switches off

CML is defined by the Philadelphia chromosome, a t(9;22) translocation that fuses the BCR gene to ABL1. The product is a constitutively active ABL1 tyrosine kinase that signals continuously through STAT5, RAS-MAPK, and PI3K-AKT. The ABL kinase domain is the druggable half of the fusion, and every approved CML small molecule except one targets the same ATP binding site in that domain. That shared dependence is exactly why a single pocket mutation can knock out an entire generation of drugs at once.

The ATP-site generations

The first five approved BCR-ABL inhibitors all compete with ATP for the catalytic cleft. They differ in which kinase conformation they prefer and how much of the resistance-mutation space they cover.

• Imatinib (Gleevec) — first-in-class, FDA approved 2001. A type-II inhibitor that binds the inactive, DFG-out conformation. Spectacular frontline durability, but a relatively narrow grip that secondary mutations escape.
• Dasatinib (Sprycel) — second generation, approved 2006. A type-I inhibitor that binds the active (DFG-in) conformation and is far more potent, covering many imatinib-resistant mutants. It is sensitive to the F317L contact mutation.
• Nilotinib (Tasigna) — second generation, approved 2007. A type-II inhibitor like imatinib but a tighter fit, with activity against many imatinib-resistant clones. It is vulnerable to Y253H, E255K/V, and F359 mutations in and around the P-loop.
• Bosutinib (Bosulif) — second generation, approved 2012. A dual SRC/ABL inhibitor used in later lines, with a distinct off-target profile.
• Ponatinib (Iclusig) — third generation, approved 2012. The first ATP-site drug deliberately engineered with a carbon-carbon triple bond linker to reach past the bulky T315I gatekeeper. Powerful and pan-mutational, but carrying a boxed warning for arterial occlusive events that shapes how aggressively it can be dosed.

Why T315I is the wall

The recurring theme in CML resistance is the gatekeeper residue threonine 315. Most ATP-site TKIs make a hydrogen bond to the threonine hydroxyl and rely on the small threonine side chain to leave room in the back of the pocket. The T315I mutation swaps in isoleucine: it removes the hydrogen-bond donor and adds a bulky hydrophobic side chain that sterically blocks imatinib, dasatinib, nilotinib, and bosutinib simultaneously. For years T315I was the mutation that ended the treatment line. Until ponatinib, it had no targeted answer, and ponatinib's cardiovascular liability meant the field still wanted a cleaner option.

The allosteric escape: asciminib

Asciminib (Scemblix, Novartis) is the structural plot twist. It does not bind the ATP site at all. Instead it occupies the myristoyl pocket, a regulatory site near the C-terminus of the ABL kinase domain that normally accepts the protein's own myristoylated N-terminal tail to hold the kinase in an autoinhibited state. BCR-ABL loses that tail in the fusion, which is part of why it stays switched on. Asciminib is a synthetic surrogate for the missing myristate: it docks into the empty pocket and clamps the kinase back into its inactive conformation. Novartis and the literature call this class a STAMP inhibitor, for Specifically Targeting the ABL Myristoyl Pocket.

Because the myristoyl pocket is physically separate from the ATP cleft, asciminib is largely indifferent to the ATP-site mutations that defeat the older drugs, including T315I. The discovery work by Wylie and colleagues established the dual-targeting rationale, and the medicinal chemistry that produced ABL001/asciminib was described by Schoepfer and colleagues. The drug first reached the clinic for heavily pretreated patients and for T315I disease.

The pivotal ASC4FIRST phase 3 trial then moved asciminib to the front line, comparing it head to head against investigator-selected standard-of-care TKIs in newly diagnosed CML in chronic phase. Asciminib met its primary endpoint with a superior major molecular response rate at 48 weeks and a favorable tolerability profile, and that readout supported its frontline approval. A drug that was designed as the answer to the resistance endgame ended up competing for first-line use, which is an unusual arc.

The structural lesson for docking

BCR-ABL is a tidy demonstration that where a ligand binds determines which mutations matter to it. The ATP-site drugs share a fate because they share a pocket. Run molecular docking with imatinib or nilotinib against wild-type ABL and you get a sensible type-II pose in the DFG-out cleft. Introduce T315I and the isoleucine side chain collides with the back of the pocket: the score degrades and the favorable hydrogen-bond geometry to residue 315 disappears. The deltadelta between wild type and T315I is the number that tells the resistance story, far more than the absolute score of either pose. An allosteric binder like asciminib lives in a different pocket entirely, so an ATP-site mutation barely touches its predicted binding mode, which is exactly the point of the design.

Try the docking yourself

Open Studio and pick ABL1 from the target catalog, then add the T315I gatekeeper mutation from the mutation chips. Dock imatinib, nilotinib, and ponatinib against wild type and against T315I and watch the first two lose the pocket while ponatinib, engineered to reach past the gatekeeper, holds on. Molecular docking makes the gatekeeper argument concrete in a way a resistance table never will.

Liganx is molecular docking online: free, browser-based, and built around exactly this mutation-versus-drug question. If you want to run molecular docking across the BCR-ABL resistance mutations without a local install, that is the fastest way to a pose you can actually inspect.

Primary sources

• Wylie AA, Schoepfer J, Jahnke W, et al. The allosteric inhibitor ABL001 enables dual targeting of BCR-ABL1. Nature 543, 733-737 (2017). doi:10.1038/nature21702
• Schoepfer J, Jahnke W, Berellini G, et al. Discovery of asciminib (ABL001), an allosteric inhibitor of the tyrosine kinase activity of BCR-ABL1. J Med Chem 61, 8120-8135 (2018). doi:10.1021/acs.jmedchem.8b01040
• Hochhaus A, Wang J, Kim DW, et al. Asciminib in newly diagnosed chronic myeloid leukemia. N Engl J Med 391, 885-896 (2024). doi:10.1056/NEJMoa2400858
• Réa D, Mauro MJ, Boquimpani C, et al. A phase 3, open-label, randomized study of asciminib versus bosutinib in heavily pretreated patients with chronic myeloid leukemia (ASCEMBL). Blood 138, 2031-2041 (2021). doi:10.1182/blood.2020009984