Therapeutics:
Breaking down resistance

from Nature Reviews Cancer

Kristine Novak

The kinase inhibitor imatinib (Glivec) can induce complete remission in patients with chronic-phase chronic myelogenous leukaemia (CML). Patients whose disease has advanced to blast crisis, however, frequently become resistant to the drug, due to mutations in the BCR–ABL kinase domain. Azam et al. have developed an in vitro screen to survey mutagenized forms of BCR–ABL, and obtained a more comprehensive picture of mutations that confer drug resistance.

Kinases typically exist in equilibrium between 'open' (active) states, or a 'closed' (autoinhibited) state. Co-crystalization studies of imatinib and the ABL kinase domain have shown that the drug achieves its specificity by trapping the kinase in the closed conformation. Most patients who become resistant to imatinib therefore harbour mutations within the BCR–ABL kinase domain. For example, mutations have been discovered that sterically hinder drug occupancy of the kinase active site, alter the phosphate-binding P loop, or influence the conformation of the loop that surrounds the active site. Different mutations confer different levels of resistance to the drug.

Azam et al. reasoned that mutations in other domains of the protein, in addition to the kinase domain, might also mediate resistance. To look for these, they randomly mutagenized the BCR–ABL gene through propagation of the gene in a bacterial strain that is deficient in DNA repair. The screen led to the isolation of 112 protein variants with distinct amino-acid substitutions in 90 residues. The kinase activity of 65 variants was confirmed, and 59 of these were found to be resistant to imatinib treatment. Only 13 of these mutations had been previously identified in patients with drug-resistant CML.

So, how do these mutations confer drug resistance? A total of 26 resistance-associated mutations were found to lie outside the kinase domain. These were instead located in the cap region (which is implicated in ABL autoinhibition), as well as in the Src homology domains 2 (SH2) and 3 (SH3), and the linker region between the SH2 and the catalytic domains. Structural modelling studies indicated that many of these mutations destabilize the closed conformation of the ABL kinase, shifting the protein equilibrium towards the open, active kinase conformation. This conformation precludes drug binding.

Analysis of these drug-resistant forms of BCR–ABL will improve the ability to predict patient drug responsiveness. For example, mutations in the SH3 domain, and some regions of the SH2 domain, generated only low levels of drug resistance. It should therefore be possible to treat patients, who are found to have mutations in these regions, with escalating doses of imatinib. The authors also discovered some P-loop mutations that confer very high levels of drug resistance. These have not yet been reported in patients, but any individuals who are found to carry these variants are unlikely to respond to imatinib. Some of the mutations were also associated with increased kinase activity and accelerated disease progression. These will be useful in determining patient prognosis.

Accompanying papers by Nagar et al. and Hantschel et al. provide a structure–function analysis of c-ABL, and a rationale for the mechanisms of many of the drug-resistant variants recovered from this screen.

References

	ORIGINAL RESEARCH PAPER Azam, M., Latek, R. R. & Daley, G. R. Mechanisms of autoinhibition and STI-571/Imatinib resistance revealed by mutagenesis of BCR–ABL. Cell 112, 831–843 (2003)
	FURTHER READING Hantschel, O. et al. A myristoyl/phosphotyrosin switch regulates c-Abl. Cell 112, 845–857 (2003) Nagar, B. et al. Structural basis for the autoinhibition of c-Abl tyrosine kinase. Cell 112, 859–871 (2003)