Molecular Analysis Of Leukemia Patients By Next Generation Sequencing Predicts Targeted Therapy Outcome

Imatinib (Gleevec/Glivec, Novartis Pharma) is an oral drug developed to fight a particular type of blood cancer, termed Chronic Myeloid Leukemia (CML), in which there is an abnormal expansion of white blood cells number.

At the molecular level, CML is caused by the formation of an aberrant chromosome, called the Philadelphia chromosome (Ph), which contains a ‘chimaeric’ gene derived from the fusion of two normal genes: BCR/ABL. This fusion gene tells the cells to divide in an uncontrolled manner, leading to leukemia. Imatinib was designed to block the activity of BCR/ABL inside the cells, in a specific way, like a key in the lock: BCR/ABL needs ATP to function, and imatinib occupies the space where ATP would bind, therefore blocking the enzyme activity.

Before imatinib, CML was a deadly disease, with a clinical course that clinicians used to divide into three phases: a long, indolent chronic phase followed by a short transition period (accelerated phase) and a final rapid acute phase termed blast crisis, which invariably led to death, unless the patient could undergo bone marrow transplant. Imatinib revolutionized the treatment and the outcome of CML: patients diagnosed in 2017 can expect a normal lifespan, while taking a pill a day, with very limited side effects.

Unfortunately, not all patients do well, even under imatinib. In some cases, the BCR/ABL gene mutates to a drug-resistant version and the disease reappears. New, more potent drugs have been developed to fight back such resistant leukemia.

In a recent paper, we analyzed a cohort of CML patients by Next Generation Sequencing. We sequenced the DNA from their leukemic cells at diagnosis and compared it with that from normal cells. We found that, in addition to the well-known BCR/ABL fusion gene, these patients had multiple other mutations in their cancer.

In order to make sense of this finding, we tried to prioritize mutations according to a novel score that we developed in the lab, the OncoScore (OS). This tool gives a score to any gene by looking in the medical literature: the more a certain gene is associated with keywords such as ‘cancer’ or ‘leukaemia’ in peer-reviewed papers, the higher its score is. We previously demonstrated that genes with OS>22 can be considered as potential oncogenes.

Interestingly, when we correlated the score of our patients mutations with their clinical response to imatinib therapy, we found that patients having higher scores had a higher probability to fail imatinib treatment, thus suggesting that additional somatic mutations in CML can modify the response to imatinib, although BCR/ABL is still the main driver of the disease. As the OS is a gene score, we compared patients carrying several mutated genes by using three ‘patient-level’ scores: the total OS (tOS), which is the sum of OncoScores of all the identified mutations in a patient; the sum of OS values considering only those genes passing the oncogene threshold (OS22 score); or by 2topOS, which is the sum of scores from the two highest-ranking mutated genes in each patient.

All patient scores yielded the same result: imatinib-resistant patients had significantly higher scores compared to responding patients, after a median of 5 years in therapy. Interestingly, the widely used Sokal score (a scoring method based on clinical parameters) was not able to discriminate the two populations, indicating that molecular profiling can be superior to traditional clinical stratification in these patients.

This study, Somatic mutations identified at diagnosis by exome sequencing can predict response to imatinib in chronic phase chronic myeloid leukemia (CML) patients was recently published in the American Journal of Hematology.