Pyruvate kinase M2 and prednisolone resistance in acute lymphoblastic leukemia

Treatment of childhood acute lymphoblastic leukemia (ALL) combines different classes of chemotherapeutic agents, such as Vinca alkaloids, anthracyclines and glucocorticoids (GCs). Although such therapy nowadays cures the majority of the patients, combination chemotherapy still fails in approximately 20%.1 Most treatment failures can be explained by resistance to antileukemic agents,2 and resistance to the glucocorticoids in particular has been shown to be related to an unfavorable event free survival.3,4 Since the glucocorticoids prednisolone and dexamethasone play a crucial role in essentially all therapy protocols, the development of strategies to reverse resistance to these agents is important to improve ALL treatment efficacy. Recently, we have demonstrated that glucocorticoid resistance in pediatric ALL is associated with increased glucose metabolism, and that inhibition of glycolysis sensitizes prednisolone-resistant ALL cells to glucocorticoids.5 Although it has been known for several decades that cancer cells shift their energy production from oxidative phosphorylation towards the less efficient glycolysis pathway (the so called Warburg effect)6 it is still not clear how tumor cells establish this altered metabolic phenotype. Christofk et al. recently showed that altered expression of the glycolytic enzyme pyruvate kinase (PK), and more specifically the switch to the alternatively spliced isoform M2 (PKM2), is responsible for the increased rate of glycolysis observed in cancer cells.7 Together, these findings suggest an upregulation of PKM2 in prednisolone resistant leukemia.

To investigate if pyruvate kinase M2 plays a role in glucocorticoid resistance in pediatric ALL, the expression of PKM2 was determined in leukemic cell samples of ALL patients and compared to PKM2 expression in normal peripheral blood and bone marrow samples. To distinguish between the different isoforms in real time quantitative PCR (RT Q-PCR), primers were designed that specifically amplify PKM2 or that recognize both isoform 1 and 2 of pyruvate kinase (PKM1/2, Figure 1A). The expression of PK isoforms was subsequently tested in different cell lines, confirming the specificity of the primer combinations (Figure 1B).

Next, the expression of the different PK isoforms was determined in normal peripheral blood lymphocytes, normal bone marrow and in leukemic samples from untreated children at initial diagnosis of ALL that were identified by an in vitro cytotoxicity assay (MTT)8 as prednisolone-resistant (n=11, LC₅₀≥150 μg/mL), or sensitive to prednisolone (n=19, LC₅₀≤0.1 μg/mL).6 In correspondence with the results of Christofk et al.,7 a significant difference (p<0.0001) was found in the expression of PKM2 between normal and ALL cells (Figure 2A). However, PKM2 transcripts were found in all patient samples and no significant difference was observed between prednisolone-resistant or prednisolone-sensitive ALL cases (Figure 2B). Similar expression patterns were observed for PKM1/2 (Figure 2C and D), although the transcript levels of PKM1/2 were about ten-fold higher than the expression levels of PKM2. Together, these data indicate that not only the M2 isoform of pyruvate kinase was present in ALL patients but also PKM1, and that the mRNA expression levels of different PKM isoforms are no indication of glucocorticoid resistance in childhood leukemia.

Figure 1.Verification of the real time PCR primer-probe specificity. (A) Graphic representation of primer design to distinguish pyruvate kinase isoforms. The primers located in exon 3 can not distinguish different PKM isoforms (PKM1/2). A primer-probe combination amplifying nt 1586–1692 (exon 9) was used to specifically amplify PKM2. Amplification efficiency was over 95% for all primer-probe combinations. Ribosomal protein S20 (RPS20) was used as a control gene for normalization. (B) Graphic representation of mRNA levels of pyruvate kinase isoform M2 (left panel) or isoform 1 and 2 (right panel) in different leukemic cell lines confirming primer-probe specificity. Expression of PK-M1/2 in Jurkat cells was set to be 100% and relative expression levels were calculated.

Figure 2.Expression of PKM isoforms in ALL patients. (A–D) Graphical representation of the expression of pyruvate kinase isoform M2 (panels a, b) or isoform M1 and M2 (panels c, d) as measured by Q-PCR. Expression of PK isoforms in normal lymphocytes versus ALL cells is depicted in panels a and c; expression of PK isoforms in prednisolone-resistant or prednisolone–sensitive patients in panels b and d. As normal samples peripheral blood (▪) or bone marrow (□) was used. Medians are indicated as horizontal lines. (E–F) Western blots representing the expression of pyruvate kinase isoforms in normal lymphocytes (panel e) or in prednisolone-resistant or prednisolone–sensitive patients (panel f). 20 μg of protein was loaded and membranes were incubated with 1:1000 diluted antibody directed against PKM1/2 or PKM2 (Cell Signaling Technology Inc., Danvers USA).

Since it has been reported that tumor cells exclusively express the M2 isoform of pyruvate kinase7,9–11 and not PKM1, we also determined the expression of the different PKM isoforms at the protein level. Western blotting was performed using antibodies that, like the primers in the Q-PCR experiments, specifically recognized PKM2 or both PKM1/2 isoforms. Blots incubated with these antibodies showed a clear band for PKM1/2, but not PKM2, in normal lymphocytes (Figure 2E). In samples from ALL patients, however, equivalent expression levels of PKM1/2 and PKM2 were present (Figure 2F). Analogous to the Q-PCR results, no significant difference in protein levels were observed between prednisolone-resistant and prednisolone-sensitive cases, suggesting that pyruvate kinase isoform M2 is not responsible for glucocorticoid resistance in childhood leukemia. Whether the difference in expression between different isoforms of pyruvate kinase that was detected between normal bone marrow and leukemic cells reflects a difference in glycolytic rate is not known, since patient cells do not grow in vitro and we can not detect glucose consumption. Thus, although pyruvate kinase might play a role in the regulation of glycolysis in childhood ALL, glucocorticoid resistance is unlikely to be caused by selective expression of PKM2.

Footnotes

• Funding: this work was financially supported by the Dutch Cancer Society (grant EMCR 2005-3313).