L-Asparaginase More Effectively Targets Leukemic Cells with Low Glycolytic Activity

Since O. Warburg first proposed aerobic glycolysis to be the main bioenergetic pathway activated in cancer cells, there has been near silence on the differences of normal and cancer metabolism. Our previous findings revealed that L-asparaginase (ASNase) triggers metabolic reprogramming of leukemic cells which is part of the adaptation process to stress caused by amino acid depletion (Hermanova et al, Leukemia 2015). We were interested if metabolic predisposition could influence the treatment response to ASNase and daunorubicine (DNR).

We studied metabolic profile of 19 leukemic cell lines covering both lymphoid (B and T cell precursor acute lymphoblastic leukemia - BCP-ALL, T-ALL) and myeloid (acute and chronic myeloid leukemia - AML, CML) lineages. First we determined the level of glycolytic activity (GA) using Seahorse Bioanalyser XF24. GA was measured as extracellular acidification rate (ECAR) and divided into four stages: basal acidification (non-glycolytic acidification), glycolysis, glycolytic capacity (maximal capacity after inhibition of oxidative phosphorylation (OXPHOS)) and glycolytic reserve (the difference between glycolytic capacity and glycolysis). GA clustered together myeloid leukemic cell lines and T-ALL with overall higher GA in comparison to BCP-ALL cell lines which gathered in the second cluster. To find the marker of GA we measured expression of genes involved in GA and showed GLUT1 (glucose transporter 1) and MCT-4 (monocarboxylate transporter 4) to correlate most significantly (r=0.634, p=0.0072; r=0.532, p=0.024) with increased glycolysis. PFKM (phospho-fructose kinase muscle) gene expression was negatively associated with basal acidification (r=-0.568, p=0.01) and glycolytic reserve (r=-0.540, p=0.02) of leukemic cells.

Another bioenergetic process, OXPHOS was measured as oxygen consumption rate (OCR). We divided OXPHOS analysis into four energetic states: basal respiration, ATP production (ATP synthesis is inhibited), maximal capacity of respiration (transport of electrons to oxygen is not limited) and spare capacity (the difference between maximal capacity and basal respiration). The OCR data in cluster analysis separated myeloid and lymphoid leukemias. Nevertheless, activity of OXPHOS positively correlated with GA in all studied cell lines showing myeloid cells to have higher oxidative metabolism compared to lymphoid cells.

Next we correlated GA and OXPHOS with IC50 of ASNase and DNR determined using MTS cytotoxic assay, representing the level of sensitivity to the drug. In general, higher GA correlated with higher IC50 of ASNase (thus lower sensitivity to ASNase): basal acidification r=0.641, p=0.003; glycolysis r=0.645, p=0.003; glycolytic capacity r=0.680, p=0.001 and glycolytic reserve r=0.790, p<0.0001. Analysing lymphoid and myeloid lineage independently we observed striking difference. While low GA correlated with higher sensitivity to ASNase in lymphoid cells (basal acidification: r=0.704, p=0.02; glycolytic capacity r=0.693, p=0.024 and glycolytic reserve r=0.648, p=0.04), in myeloid cells correlation with higher sensitivity to ASNase was observed only for the glycolytic reserve (r=0.775; p=0.0211). There was no such correlation between IC50 of ASNase and OXPHOS. Sensitivity to DNR correlated neither with GA nor OXPHOS across the entire sample set. However, in lymphoid cell lines there was a correlation of lower sensitivity to DNR with enhanced GA and OXPHOS (basal acidification r=0.848, p=0.001; glycolytic capacity r=0.796, p=0.004 and glycolytic reserve r=0.897, p=0.0001; basal respiration r=0.747, p=0.01; ATP production r=0.716, p=0.02; maximal capacity r=0.742, p=0.01).

In conclusion, we found significant difference between the key metabolic activities in myeloid and BCP and T lymphoid malignant cells. Our data suggest association between basal glycolytic state of leukemic cells and response to ASNase and DNR. Lower GA of lymphoid cells (in particular BCP) is associated with higher sensitivity to ASNase and DNR. In myeloid cells GA does not correlate with sensitivity to ASNase and DNR probably due to overall high GA. We assume that activation of glycolytic reserve under metabolic stress can affect the sensitivity to ASNase in myeloid cells. The effect of ASNase and DNR seems to be related to metabolic state of leukemic cells only in lymphoid leukemic cells.

Supported by IGA NT1249, 15-28848A, UNCE204012.