Abstract

We previously reported that imatinib decreased glucose uptake in the BCR-ABL+ cells in a dose- and timedependent fashion by switching from glycolysis to Krebs cycle metabolism and improving the cell energy state. The goal of this study was to evaluate metabolic aspects of imatinib resistance in paired imatinib-sensitive and resistant clones of two CML cell lines, K562 and LAMA84. LAMA84-r cells are resistant due to BCR-ABL and P-glycoprotein overexpression, the mechanism of resistancy in K562-r cells is still unknown (may be implicating other tyrosine kinases). Resistant clones were grown in imatinib containing medium (K562-r: 5μM; LAMA84-r: 1μM). Sensitive cells were incubated with 1μM imatinib for 24h. [1-13C] glucose was added to all cells for the last 4h. Magnetic resonance spectroscopy (MRS) was performed on cell extracts to evaluate glucose metabolism (1H, 13C) and energy state (31P). In imatinib treated sensitive cells mitochondrial glucose metabolism was slightly improved (C4-glutamate signal increased to 137% and 114% vs. untreated K562-s and LAMA84-s). No significant changes in glycolysis rate or glucose uptake were seen. In resistant K562-r cells mitochondrial activity was significantly lower (C4-glutamate decreased to 47%, p<0.001) with decreased citrate (p<0.05) and pyruvate (p<0.05) concentrations and decreased NTP/NDP-ratios (45%, vs. K562-s p<0.05). The de novo13C-lactate production decreased in K562-r cells (46%, p<0.05), but increased in the media (168%, p<0.05, all compared to K562-s), in order to keep the intracellular pH stable. K562-r showed a significant increase in glucose uptake vs. sensitive cells (17.2 compared to 13.6 mmol/l/g, p<0.05). Even though LAMA84-r showed the same metabolic pattern, the changes in this cell line were not statistically significant. The pronounced decrease in phosphodiester concentrations, markers for membrane degradation, clearly distinguished resistant from the sensitive cells (21%, p<0.005 (K562) and 55%, p<0.05 (LAMA84)). In LAMA84-r cells concentration of phosphocholine, a marker for membrane biosynthesis, was increased to 293% of LAMA84-s (p<0.001). Preliminary GC-MS studies on K562-r cells showed that imatinib is unable to inhibit non-oxidative nucleic acid ribose and deoxyribose synthesis. A significantly increased activity of transketolase (p<0.05) the enzyme responsible for non-oxidative ribose synthesis in the pentose cycle, may be a putative mechanism for this salvage reaction. In summary, despite their different mechanisms of resistance, K562-r and LAMA84-r cells showed a metabolic profile of the resistant phenotype with same directional changes compared to their sensitive counterparts. Mitochondrial metabolism and energy state were decreased significantly in K562-r (p<0.05), while non-oxidative metabolism was not affected. These changes were present but non-significant in LAMA84-r. We therefore conclude that the failure of imatinib in inhibiting cytosolic glycolysis and improving mitochondrial metabolism, as well as in restricting the use of glucose carbons for de novo nucleic acid synthesis could serve as a metabolic signature of imatinib resistance. This can be readily measured by MRS and may be used as a clinical endpoint for early detection of treatment failure.

Author notes

Corresponding author