The Pill That Could · Resistance
Resistance, part 2
If this was happening, you would expect to see, in resistant cells, mutations in the BCR-ABL gene itself — specifically in the part of the gene encoding the kinase domain, and especially in the parts of the kinase domain that contacted the drug.
Sawyers, with a postdoctoral fellow named Mercedes Gorre and other members of his lab, took up the question. They obtained samples of blast cells from a series of CML patients who had been on imatinib and relapsed. They sequenced the BCR-ABL gene in each sample, looking specifically at the kinase domain.
What they found, published in Science in 2001, was that most of the resistant patients had mutations.
The mutations were not all the same. Different patients had different mutations, in different places in the kinase domain. But they all clustered in a region of the protein consistent with the drug's binding site — the parts of the kinase that, in the three-dimensional structure, would be in contact with imatinib. One mutation appeared in multiple unrelated patients, suggesting it was a particularly favored escape route: a substitution of an isoleucine residue at position 315 of the ABL protein for a threonine, written in the standard one-letter code as T315I. (The T is the original threonine, the 315 is the position number, the I is the isoleucine that replaced it.) Other mutations, occurring at other residues that contacted the drug, included Y253F, E255K, M351T, and others. Each of them, at the level of the protein, made some small change in the shape of the binding site. Each of them, at the level of the cell, made imatinib bind less well.
The cancer was, in short, doing exactly what evolution would predict. It was acquiring point mutations in the gene that the drug targeted, and the cells with the right mutations were the ones that survived.
* * *
It is worth pausing on T315I, because it is going to come back several times in this chapter, and because it is the place where the resistance story turns into a story about chemistry as much as biology.
The T315 residue — the threonine at position 315 of the ABL protein — is in a part of the kinase domain called the gatekeeper position. The name is descriptive. The gatekeeper is the residue that sits at the back of the ATP pocket, helping to determine which molecules can fit deep into the pocket and which can't. In ABL, as in many tyrosine kinases, the gatekeeper is a small amino acid — threonine has a relatively short side chain — which leaves room for an ATP-competitive inhibitor to extend a chemical group past the gatekeeper into a pocket beyond it. Imatinib makes specific use of this geometry. A small projection of the imatinib molecule slips past the threonine gatekeeper and contacts a region of the kinase that, in less selectively targetable kinases, would be inaccessible. The threonine, by virtue of being small, lets the drug in.
The T315I mutation replaces threonine with isoleucine. Isoleucine has a longer, bulkier side chain. Where the small threonine had left room for the drug to extend past it, the bulky isoleucine fills the space. The drug can no longer slip past the gatekeeper. It can no longer bind tightly. The kinase is now invisible to imatinib, in the sense that the drug can't grab hold of it.
This is, at the level of one amino acid, why a single point mutation can make a cell completely resistant to a drug that, in cells without the mutation, works almost perfectly. It is also, at the level of strategy, why T315I is harder to address than other mutations. The other resistance mutations — the ones at Y253, E255, M351, and so on — mostly disrupt imatinib binding through subtler means: shifts in the conformation the kinase prefers, slight changes in the contacts the drug makes elsewhere in the pocket. A second-generation inhibitor that bound the kinase a little differently, or a little more tightly, could often work around them. T315I was different. The gatekeeper substitution closed off a part of the binding site that imatinib actively used. Designing a drug that worked despite a closed gatekeeper required something fundamentally different from imatinib, not just a better-tuned version.
We will come back to T315I.
* * *
What followed Sawyers's 2001 paper was a flurry of work, in academic labs and pharmaceutical companies, to make second-generation BCR-ABL inhibitors that could overcome resistance.
Two compounds emerged most prominently. The first was a Bristol-Myers Squibb compound originally numbered BMS-354825 and eventually named dasatinib. It was a molecule with a different chemical scaffold from imatinib — it inhibited BCR-ABL through a slightly different mode of binding, and it was, by orders of magnitude, more potent than imatinib against the wild-type kinase. It also inhibited most of the resistance mutations that had emerged in imatinib-treated patients. It did not inhibit T315I.
The second was a Novartis compound, AMN107, eventually named nilotinib. Novartis, having developed imatinib, had a strong interest in extending the franchise; nilotinib was, in a sense, an explicit second-generation imatinib. The chemists at Novartis had taken the imatinib structure and modified it carefully, addressing some of the chemical issues that had limited imatinib's potency and improving its activity against most of the common resistance mutations. Nilotinib, like dasatinib, did not inhibit T315I.
Both drugs were approved, in 2006 and 2007 respectively, for CML patients who had become resistant to or intolerant of imatinib. They worked. Most patients with imatinib-resistant CML who were not carrying the T315I mutation could be returned to remission with one of the second-generation drugs. The arsenal of effective CML therapies, which in 2000 had consisted of one drug, expanded to three by 2007. Patients who relapsed on imatinib could be switched to dasatinib or nilotinib; if they relapsed again, they could often be switched to the other; many of them could be kept in cytogenetic remission through a sequence of these drugs for years.
All 13 chapters of The Pill That Could · Library · All courses