Flow Cytometry Detection of Intra-Cellular Tyrosine Kinase Inhibitors (TKI) Showed Variable Uptake in CML CD34+ Cells

Abstract 2747 Introduction Despite the major benefit of TKI in the treatment of Chronic Myeloid Leukemia (CML), patient response is heterogeneous and it is generally accepted that residual disease and relapse are due to persistent CML cells, considered as leukemic stem cells. Their resistance has been related to lower TKI uptake. The amount of drug penetrating the targeted cells is most likely a major parameter of targeted therapy efficacy since it is essential that the therapeutic molecule be as close as possible to the target molecule. We developed a flow cytometry technique to analyze primary cells. Method To evaluate intracellular imatinib (ICIM) uptake, we developed a patented method based on natural UV fluorescence related to chemical structure. Consequently, since the difference in UV fluorescence units between treated and control cells is proportional to the amount of intra-cellular drugs, we validated this method after incubating K562 and KCL22 cell lines with TKI. The flow cytometry technique was standardized by using Flow-Check Fluorosphere calibrated beads immediately before, and at the end of, each series of analyses with a Coulter Epics Elite™ flow cytometer (Beckman Coulter) equipped with an Innova I90C-4 UV laser (Coherent). Then we analyzed primary blood cells from CML patients in chronic phase before any treatment. After lysis of erythrocytes, nucleated cells were incubated at 1.106 cells/ml with different doses of imatinib (IMA) (n=22), Nilotinib (NIL) (n=20) and Dasatinib (DAS) (n=20) at different times. Whenever possible, CML stem cells were analyzed using CD34-FITC staining. Results In preliminary assays, we checked that there was a significant correlation between additional fluorescence measured by flow cytometry and the amount quantified by physico-chemical analysis after lysing a known number of cells (n=57, r2=0.73, p<0.001), which enabled us to convert UV fluorescence into pg of IMA per cell. Then we confirmed that IMI rapidly penetrated K562 and KCL22 cells (from 5 minutes of incubation) and reached a stable influx in viable cells from 1 hour (T1h). We chose this incubation time for further experiments. Similarly, we choose T2h for second generation TKI. We observed a dose-dependent accumulation in the two cells lines, but with differences at the lowest extra-cellular concentrations (1–5 μM) and not correlated with any membrane pump expression (OCT-1, ABCG2, ABCB1 and ABCC1). ICIM at T1h was correlated with cell sensitivity to IM at T24h expressed by the proportion of dead cells (r2=0.93 and 0.88 for K562 and KCL22 cells, respectively). We then applied our method to primary CML blood cells in comparison with normal blood cells. TKI penetrated all cell subsets, but amounts varied depending on cell sizes (FS/SS characteristics). The first data obtained with IM showed ICIM levels in CML cells that were relatively heterogeneous from one patient to another, ranging from 0.9 to 4 pg/cell for an extracellular concentration of 5 μM, i.e. a higher concentration (x 300) than in culture medium. The ability of the granulocyte cell lineage to store IMA was related to the Sokal prognostic index (p=0.05). We detected variable ICIM levels in CML CD34+ cells from 10/16 patients (0.04–0.7 pg/cell) and no signal for 6/16 patients. Surprisingly, the ability of CD34+ to store second generation TKIs is variable and not necessarily correlated to IMA uptake. Discussion We developed a simple, rapid flow cytometry method directly applicable to primary cells and requiring only few cells which makes it possible to identify target cell subsets, such as CML stem cells. The strong correlation between the ICIM amount and the sensitivity of CML cell lines to TKIs validated the method and suggested that ICIM could be a relevant biomarker for predicting the sensitivity of the CML clone. In our CML series, we observed striking inter-patient variability of the capacity of primary CML cells to store TKI. A correlation with the Sokal score suggests possible predictive value with regard to in vivo CML response to IMA, which could be taken into account when choosing TKI for first-line therapy. Furthermore, we observed marked heterogeneity between CML CD34+ cells for storing TKI that could partially explain the heterogeneity of in vivo response. The relationship between the ability of untreated CML CD34+ cells to store TKI and complete molecular response has to be established. Disclosures: No relevant conflicts of interest to declare.