Rotating between ponatinib and imatinib temporarily increases the efficacy of imatinib, as shown in a model of chronic myeloid leukemia
Chronic myeloid leukemia (CML) has defined the paradigm for targeted cancer treatment, following the resounding success of imatinib in the early 2000s. These targeted therapies consist of tyrosine kinase (TKI) inhibitors, which eliminate the aberrant kinase activity of Bcr-Abl1. However, the precision with which targeted therapies work means that even small changes can reduce their effect, making them vulnerable to the clonal evolution present in all forms of cancer.1,2,3. Unfortunately, even during the first trials, cases of patients who became resistant began to arrive.4. Since then, similar accounts exist for all drugs used in the treatment of CML5. Unlike most forms of cancer, CML has a single molecular driver, caused by a chromosomal translocation, that essentially only exists in cancer cells: Bcr-Abl16. For this reason, CML responds so well to targeted therapy, because the sole cause of the malignancy can be effectively suppressed by targeting only Bcr-Abl17. On the other hand, the inhibition of Bcr-Abl1 exerts a strong evolutionary pressure on cancer, which is at the origin of the development of dozens of different resistance mutations. As a result, a significant fraction of CML patients relapse with some adaptation to drug resistance.
Most commonly, resistance occurs through mutations in the kinase domain (KD) of Bcr-Abl1, impairing drug binding and restoring oncogenic signaling8. Sensitive diagnosis of mutations is paramount for the success of targeted therapy in CML 9. Often these mutations offer specific protection against one or a few drugs (Fig. 1), the most problematic being the well-known gatekeeper mutation T315I. This mutation offers a high degree of resistance to all drugs except ponatinibten.
While mutations and increased expression of Bcr-Abl1 are the primary Bcr-Abl1-dependent resistance mechanisms, other changes in cancer cells may have protective effects. Many TKIs are substrates of drug efflux transporters, particularly ABCB1 and ABCG2; an increase in their activity has been associated with resistance12. It should be noted that this does not appear to affect all TKIs, and it has been suggested that TKIs that are unaffected by such transporters may be better first-choice drugs. 13. Another possibility is to replace and/or complement the role of Bcr-Abl1 as a singular oncogenic motor. Most commonly, this involves abnormal Src activity. Apparently some second-generation TKIs are also active against Src (dasatinib, bosutinib, and ponatinib), which may provide an additional protective effect14.
Several approaches have been suggested to limit the scope of drug resistance. The most common approach is to wait for the first signs of relapse and then attempt to select a drug that should be effective against any resistance adaptation that may be present. Typically, Bcr-Abl1 KD sequencing is done to identify mutations and select a drug accordingly, which can be beneficial15. This methodology is currently applied clinically, but not universally.16. Another approach to reduce the risk of drug resistance is to combine drugs11,17,18, where two or more drugs are used at the same time. By selecting drugs with complementary mechanisms and non-overlapping resistance adaptations, they can be made both robust and effective. In particular, the new inhibitor asciminib shows promise in this regard.11,18,19. However, drug combinations may not be tolerable or may have unpredictable negative interactions. Yet another strategy is to use drug rotations where therapy alternates between two or more drugs. The idea is that periodic switching from one drug to another should reduce the incidence of resistance, since a trait that confers resistance to drug A might not protect against drug B and vice versa. This strategy may be considered less beneficial than combination therapy, but does not suffer from the same risk of negative interactions20.21. We have previously predicted that beneficial drug rotations may exist in CML using currently approved drugs22 with computer modeling. For this to work, it is essential that the drugs have a different set of resistance mutations (eg, Fig. 1). Drug rotations involving ponatinib, which is the only drug effective against the T315I mutation (in the simulation study), were expected to have the greatest effect.
To estimate the viability of a drug rotation approach as suggested in our previous study22, we exposed KCL-22 to a drug rotation of imatinib and ponatinib (K562 was also tested, but did not survive therapy despite multiple trials). KCL-22 cells appear to become resistant easily when cultured with a TKI and do not require long-term exposure23. Cells were cultured in multiple replicates on a multiwell plate in the presence of imatinib or ponatinib. Drug concentrations were chosen such that cells would grow at about half their normal rate. This made it possible to maintain the cell densities of a normal culture protocol for each of the cell lines, and to monitor their growth by periodic counting. Every six days, cells were washed to remove the previous inhibitor (or simulate washing stress for control groups), and reseeded with a new inhibitor while maintaining an appropriate population density, in a new plate. This was repeated six times, generating a population-sized timeline as part of a drug rotation protocol or single-drug control treatment.