The Pill That Could · Resistance

Resistance, part 3

There was a parallel question, opened up by the second-generation drugs, of whether they should be used not just after imatinib but instead of it, as initial therapy. The argument for using them first was that they produced faster and deeper molecular responses than imatinib did, with the implication that fewer patients would progress to advanced disease and fewer resistance mutations would have time to emerge. The argument for using imatinib first was that imatinib had longer follow-up data, well-characterized side effects, and — by the time the question became live in the late 2000s — the prospect of going generic, which would dramatically lower the cost of first-line therapy. Both nilotinib and dasatinib were approved for first-line use in the early 2010s, but the field's practice settled into a roughly even split: some physicians began with imatinib, some began with one of the second-generation drugs, and the choice depended on patient factors and physician preference more than on a clear superiority of either approach for most patients. We will come back to the cost question in chapter eleven.

But there was still T315I, the gatekeeper mutation, against which neither dasatinib nor nilotinib was effective.

* * *

T315I is, by some estimates, present in fifteen to twenty percent of imatinib-resistant patients — not the majority, but not a small minority either. For these patients, dasatinib and nilotinib were as ineffective as imatinib. The cancer had taken the route through the gatekeeper, and the only way around the gatekeeper was a fundamentally different molecule.

Several pharmaceutical and academic teams worked on the problem through the 2000s. The chemistry was difficult. The constraints on a molecule that could bind a closed-gatekeeper kinase were severe; the chemists had to find a way for the drug to engage the kinase without using the part of the binding site the gatekeeper normally left open.

A small biotechnology company called ARIAD Pharmaceuticals, based in Cambridge, Massachusetts, eventually produced one. Their compound, originally numbered AP24534 and later named ponatinib, was designed specifically to inhibit T315I-mutant BCR-ABL. It used a chemical strategy quite different from imatinib's: rather than slipping past the gatekeeper, it bound in a configuration that didn't depend on getting past it at all. It was, in the test tube, potent against essentially every BCR-ABL mutation that had been identified in imatinib-resistant patients, including T315I.

Ponatinib was approved in 2012, for patients with CML and Philadelphia-positive acute lymphoblastic leukemia who carried the T315I mutation or whose disease was resistant to other available kinase inhibitors. It was, for those patients, often the difference between effective therapy and no effective therapy.

It came with a problem the earlier drugs had not had. Ponatinib turned out to cause an unexpectedly high rate of arterial thromboses — blood clots in arteries, leading to heart attacks, strokes, and limb-threatening blockages — in patients taking it. The mechanism was poorly understood; it was not obviously related to BCR-ABL inhibition itself, but rather to off-target effects of the molecule on other targets in the cardiovascular system. The frequency of these events was high enough that the FDA, in 2013, briefly suspended the drug's marketing while the data were reviewed. Ponatinib was eventually returned to the market, but with a black-box warning, with revised dose recommendations (lower doses than originally approved, with the dose lowered further once a patient achieved a satisfactory response), and with a clear understanding among prescribing oncologists that the drug was for patients whose situation was bad enough to justify accepting the cardiovascular risk.

This is, in itself, a piece of the resistance story worth attending to. The drug that closed the gatekeeper escape route for imatinib was a more dangerous drug than its predecessors. The molecular ingenuity required to build it had come at a cost in tolerability. The arc of CML treatment, which had begun with the cleanest of clean targeted therapies, had bent toward something messier as the cancer's evolutionary repertoire forced the field to use more aggressive molecules to keep up.

* * *

There is, in 2026, a fourth drug specifically designed for T315I, called asciminib, that takes a different approach again.

Asciminib was developed by Novartis and approved in 2021. It is what chemists call an allosteric inhibitor: rather than binding the ATP pocket the way imatinib, dasatinib, nilotinib, and ponatinib all do, it binds a different site on the kinase — a site called the myristoyl pocket, which is the pocket that, in the normal regulation of ABL, accepts a small fatty-acid tag and uses the binding to keep the kinase folded in its inactive conformation. BCR-ABL, lacking the regulatory machinery that uses this pocket, has the pocket sitting empty. Asciminib slots into the empty pocket and locks the kinase in the inactive conformation — turning the same regulatory mechanism that keeps normal ABL quiet against BCR-ABL, which had escaped the regulation by being a fusion. Because asciminib doesn't bind in the ATP pocket, it isn't affected by the gatekeeper mutation that defeats the ATP-competitive drugs. Because it engages a regulatory site that doesn't exist in most other kinases, it is unusually selective — even more so than imatinib was. Its safety profile, in the trials that led to its approval, has been better than ponatinib's.

Asciminib is the kind of drug that becomes possible only after years of structural work on a target. The myristoyl pocket was first appreciated as a regulatory site in ABL in the 1990s. The realization that it could be drugged, and the chemistry to drug it, took the better part of two decades after that. Asciminib's approval is, in 2026, one of the more striking examples in oncology of medicinal chemistry catching up with structural biology — of the ability to design a drug that exploits a feature of a protein nobody knew how to use, twenty years earlier, when the first inhibitor of the same protein was approved.

All 13 chapters of The Pill That Could · Library · All courses