Distinct Activities of Glycolytic Enzymes Identify Chronic Lymphocytic Leukemia Patients with a more Aggressive Course and Resistance to Chemo-Immunotherapy

Abstract

A higher capacity to grow under hypoxic conditions can lead to a more aggressive behavior of tumor cells. Determining tumor activity under hypoxia may identify chronic lymphocytic leukemia (CLL) with aggressive clinical course and predict response to chemo-immunotherapy (CIT). A metabolic score was generated by determining pyruvate kinase and lactate dehydrogenase, key enzymes of glycolysis, ex vivo in primary CLL samples under normoxic and hypoxic conditions. This score was further correlated with clinical endpoints and response to CIT in 96 CLL patients. 45 patients were classified as metabolic high risk (HR), 51 as low risk (LR). Treatment-free survival (TFS) was significantly shorter in HR patients (median 394 vs 723 days, p = .021). 15 HR patients and 14 LR patients received CIT after sample acquisition. HR patients had a significantly shorter progression-free survival after treatment compared to LR patients (median 216 days vs not reached, p = .008). Multivariate analysis evaluating age, IGHV, TP53 deletion or mutation and 11q22-23 deletion besides the capacity of tumor cells to grow under severe hypoxic conditions identified the metabolic profile as the strongest independent risk factor for shorter TFS (hazard ratio 2.37, p = .011). The metabolic risk can provide prognostic and predictive information complementary to genetic biomarkers and identify patients who might benefit from alternative treatment approaches.

Keywords: CLL; High-risk; LDH; Metabolism; PK M2.

Fig. 1

Glycolytic enzymes and glucose flux in Mec-1 cells. 10 [4] Mec-1 cells were treated with 10 μM P-M2tide or 10 μM DASA, respectively (both with no cytotoxicity towards Mec-1 cells), and with fludarabine (simultaneously, under hypoxia for 24 h, n = 4; A). Viability was defined as % luminescence of the untreated control. Lactate efflux after 24 h was measured using 10⁶ Mec-1 cells (n = 2; B). 5-³H-glucose turn-over was assessed after treatment with PM2-tide or DASA (under hypoxia, 24 h, n = 3; C). For qRT-PCR fold change of expression (compared to untreated control) of glycolytic Iso−/enzymes Hexokinase (HK), Glucose-6-phhosphate isomerase (GPI), Phosphofructokinase (PFKL, two different primers), Aldolase A/B/C (ALDOA, ALDOB and ALDOC each two different primers), Triose-phosphate isomerase (TPI), Phosphoglycerate mutase (PGM), Enolase (ENO), Pyruvate kinase (PKLR two different primers (PKLR was not detectable with primer #1); PKM2) and Lactate dehydrogenase (LDHA, LDHC with two primers (#1 and #2) and repetitive testing) was measured in untreated cells and after treatment with P-M2tide or DASA (under hypoxia, 24 h, n = 3; D). A minimum of 2-fold change (relative quantification of more than two or <0.5; dashed lines) was considered significant. Error bars are indicating standard deviation. Statistical significance was calculated with a two-way t-test. Significance is represented as * for p-values <.05.

Fig. 2

Activity of phosphofructokinase and hexokinase in normoxia and hypoxia after pharmacological modulation of PK M2. Homogenates from 10⁶ Mec-1 cells used for the experiments shown in Fig. 1B and Fig. 1D were analyzed for specific enzyme activities (A, B). There were no significant changes in activities.

Fig. 3

Treatment-free survival according to dichotomized metabolic score (D-MS). Analysis of all CLL patients (HR n = 45, LR n = 51; A), as well as subgroups that are TP53 deleted or mutated (HR n = 8, LR n = 10; B), with del11q22–23 (HR n = 8, LR n = 6; C), IGHV-M (HR n = 17, LR n = 22; D) and IGHV-U (HR n = 18, LR n = 16; E).

Fig. 4

Progression-free survival after treatment with chemo-immunotherapy according to dichotomized metabolic score (D-MS). Analysis of all CLL patients treated with chemo-immunotherapy (HR n = 15, LR n = 14; A). Subgroup analysis after exclusion of patients that are TP53 deleted or mutated (HR n = 11, LR n = 9, B). Subgroup analysis of patients treated with the combination of bendamustine with rituximab (BR; HR n = 9, LR n = 8; C). Based on the results of CLL patient sample analysis, the cut-off of the test value |MS-2| was variated in the sense of an ROC analysis (HR n = 15, LR n = 14). It was ascertained that the sensitivity and the specificity were similar at a value of 0.3 correlating about 75% (D). The discriminatory performance of the test value |MS-2| is shown in the ROC curve with an area under curve of 0.77 (HR n = 15, LR n = 14; E). PFS was determined in CLL patients who were treated with BCRi after sampling (n = 14, ibrutinib/idelalisib HR n = 5/2, LR n = 3/4; F).