The metabolic function of pyruvate kinase M2 regulates reactive oxygen species production and microbial killing by neutrophils

Abstract

Neutrophils rely predominantly on glycolytic metabolism for their biological functions, including reactive oxygen species (ROS) production. Although pyruvate kinase M2 (PKM2) is a glycolytic enzyme known to be involved in metabolic reprogramming and gene transcription in many immune cell types, its role in neutrophils remains poorly understood. Here, we report that PKM2 regulates ROS production and microbial killing by neutrophils. Zymosan-activated neutrophils showed increased cytoplasmic expression of PKM2. Pharmacological inhibition or genetic deficiency of PKM2 in neutrophils reduced ROS production and Staphylococcus aureus killing in vitro. In addition, this also resulted in phosphoenolpyruvate (PEP) accumulation and decreased dihydroxyacetone phosphate (DHAP) production, which is required for de novo synthesis of diacylglycerol (DAG) from glycolysis. In vivo, PKM2 deficiency in myeloid cells impaired the control of infection with Staphylococcus aureus. Our results fill the gap in the current knowledge of the importance of lower glycolysis for ROS production in neutrophils, highlighting the role of PKM2 in regulating the DHAP and DAG synthesis to promote ROS production in neutrophils.

Fig. 1. PKM2 expression increases in activated neutrophils.

a Pkm2 expression analysis in myeloid cells from the Immunological Genome Project (ImmGen) bulk-population RNA-seq database. b Immunoblot analysis of PKM2 expression in BM-isolated neutrophils from wild-type (Pkm2^fl/fl) and PKM2-deficient (Pkm2^∆Lyz2) mice. Cells were activated with Zy/op (100 µg/mL) or medium only (Ctrl) for 6 h. β-actin was used as the loading control. Representative of three independent experiments. c Immunoblot analysis of PKM2 conformational states in murine neutrophils after activation with Zy/op for 6 h using DSS cross-linking assay. Representative of two independent experiments. d Confocal immunofluorescence analysis of Zy/op-activated murine neutrophils stained for PKM2 (green) and DAPI (blue). The scale bar indicates 5 μm. Representative of three independent experiments. e Immunoblot analysis of PKM2 expression in blood-isolated neutrophils from three healthy volunteers. Cells were activated with Zy/op for 6 h. β-actin was used as the loading control. Representative of three independent experiments. f Confocal immunofluorescence analysis of Zy/op-activated human neutrophils stained for PKM2 (green) and DAPI (blue). The scale bar indicates 5 μm. Representative of three independent experiments. Clip art provided by Biorender. Source data are provided in the Source data file.

Fig. 2. PKM2 deficiency impairs ROS production and microbial killing by neutrophils.

a Phagocytosis of Zy/op-FITC (100 µg/mL) by wild-type (Pkm2^fl/fl) and PKM2-deficient (Pkm2^∆Lyz2) neutrophils. Representative dot plots, histograms and bar graphs showing the frequency and the median fluorescence intensity (MFI) of phagocytic cells. Gating strategies are shown in Supplementary Fig. 5a. n = 3 mice per group, representative of two independent experiments. b Kinetic measurement of ROS production by wild-type and PKM2-deficient neutrophils activated with Zy/op. Representative time-response graph and area under the curve (AUC) graph bar. n = 5 mice per group, representative of five independent experiments. c ROS production by wild-type and PKM2-deficient neutrophils exposed to S. aureus (2 ×10⁶) opsonized with serum. n = 3 mice per group, representative of two independent experiments. d Quantification of viable S. aureus recovered from lysates of wild-type and PKM2-deficient neutrophils after 2 h. n = 3 mice per group, representative of three independent experiments. e ROS production by human neutrophils pre-treated with oxalate (3 mM) for 1 h and then activated with Zy/op. Representative time-response graph and AUC graph bar. n = 4 donors in four independent experiments. f ROS production by human neutrophils pre-treated with TEPP-46 (30 µM) for 1 h and then activated with Zy/op. Representative time-response graph and AUC graph bar. n = 4 donors in four independent experiments. g Number of viable S. aureus recovered after 2 h from lysates of human neutrophils pre-treated with oxalate or TEPP-46 for 1 h. n = 5 donors in three independent experiments. Error bars are mean ± SEM. p values were determined by two-tailed unpaired Student’s t-test (a, d) or one-way ANOVA followed by Tukey’s post hoc test (b, c, e–g). Clip art provided by Biorender. Source data are provided in the Source data file.

Fig. 3. Aerobic glycolysis is essential for ROS production in activated neutrophils.

a Schematic representation of the two simultaneous assays pooled together in the graphs of Fig. 3b, c. b Kinetic profile of ECAR and ROS and c, OCR and ROS production in neutrophils activated with Zy/op (100 µg/mL). n = 5 mice per group, representative of two independent experiments. d Schematic of the steps where the drugs inhibit glycolysis. e ROS production by neutrophils pre-treated with 6-AN (3 mM) for 1 h and then activated with Zy/op for 1 h. n = 3 mice per group, representative of three independent experiments. f NADPH and total NADP production by neutrophils activated with Zy/op for 30 min. n = 4 mice per group, representative of two independent experiments. g gp91-phox mRNA expression normalised to Gapdh in neutrophils activated with Zy/op for 3 h. n = 5 mice per group, representative of two independent experiments. h Immunoblot analysis of gp91-phox and p47-phox expression in neutrophils from wild-type and PKM2-deficient mice. β-actin was used as the loading control. Representative of two independent experiments. i ROS production by wild-type murine (left) or human (right) neutrophils pre-treated with 2-DG (3 mM), 3PO (10 µM), or oxalate (3 mM) for 1 h, then activated with Zy/op for 1 h. n = 5 mice per group in five independent experiments and n = 4 donors in four independent experiments, respectively. j ROS production by neutrophils pre-treated with a PLC inhibitor (iPLC, 1 µM) for 1 h, then activated with Zy/op for 1 h. n = 5 mice per group, representative of two independent experiments. k Representative real-time ROS production in wild-type and human neutrophils activated with Zy/op. ROS production was measured in response to glucose, 3PO, and iPLC at the indicated time points. n = 3 mice per group, representative of two independent experiments and n = 4 donors in four independent experiments, respectively. Error bars are mean ± SEM. p values were determined by two-tailed unpaired Student’s t-test (f) or one-way ANOVA followed by Tukey’s post hoc test (b, c, e, g, i–k). Clip art provided by Biorender. Source data are provided in the Source data file.

Fig. 4. Metabolic activity of PKM2 regulates ROS production in neutrophils.

a GLUT1 expression in wild-type (Pkm2^fl/fl) and PKM2-deficient (Pkm2^∆Lyz2) activated with Zy/op (100 µg/mL) for 1 h. Representative histogram and graph bar showing the GLUT1 frequency and median fluorescence intensity (MFI). Gating strategies are shown in Supplementary Fig. 5a. n = 6 mice per group in two independent experiments. b Glucose uptake in wild-type and PKM2-deficient neutrophils activated with Zy/op for 30 min. Representative histogram and graph bar showing frequency and MFI for 2-NBDG. Gating strategies are shown in Supplementary Fig. 5a. n = 6 mice per group in two independent experiments. c Pyruvate Kinase (PK) activity in neutrophils activated with Zy/op for 1 h. n = 5 mice per group in two independent experiments. d Pyruvate production by neutrophils activated with Zy/op for 1 h. n = 4 mice per group in two independent experiments. e Lactate production by wild-type and PKM2-deficient mouse neutrophils and by human neutrophils pre-treated with oxalate (3 mM) for 1 h and activated with Zy/op for 1 h. n = 5 mice per group in two independent experiments and n = 5 donors in five independent experiments. f Glycolytic metabolites from ¹³C-glucose tracing in wild-type and PKM2-deficient neutrophils activated with Zy/op for 1 h. The y-axis refers to the height of the chromatography peak obtained by LC-MS. n = 3 mice per group in three independent experiments. Error bars are mean ± SEM. p values were determined by one-way ANOVA followed by Tukey’s post hoc test. Clip art provided by Biorender. Source data are provided in the Source data file.

Fig. 5. PKM2 modulates ROS production in neutrophils by a TPI-dependent mechanism.

a Schematic of glycolysis inhibition by various drugs. b ROS production by neutrophils treated with 5-PDR (30 µM) or propranolol (30 µM) for 1 h and activated with Zy/op for 1 h. n = 3 mice per group, representative of three independent experiments. c Immunoblot analysis of p47-phox phosphorylation in wild-type (Pkm2^fl/fl) and PKM2-deficient (Pkm2^∆Lyz2) neutrophils activated with Zy/op for 30 min. Representative of two independent experiments. d ROS production by neutrophils activated with Zy/op or PMA (50 nM) for 1 h. n = 6 mice per group, representative of three independent experiments. e ROS production by murine and human neutrophils pre-treated with oxalate (3 mM) for 1 h and activated with Zy/op for 1 h in the presence of OAG (10 µM). n = 3 mice per group in three independent experiments and n = 4 donors in four independent experiments, respectively. f ROS production by neutrophils pre-treated with oxalate or PEP (1 mM) for 1 h and activated with Zy/op for 1 h. n = 3 mice per group, representative of three independent experiments. g DHAP production by neutrophils pre-treated with PEP for 1 h and activated with Zy/op for 1 h. n = 3 mice per group, representative of two independent experiments. h G3P production by neutrophils pre-treated with PEP and activated with Zy/op. n = 4 mice per group, representative of two independent experiments. i TPI activity in neutrophils pre-treated with PEP and activated with Zy/op. n = 3 mice per group, representative of two independent experiments. j Immunoblot analysis of p47-phox phosphorylation in wild-type neutrophils pre-treated with oxalate, PEP, 5-PDR or propranolol for 1 h and activated with Zy/op for 30 min. Representative of two independent experiments. k TPI activity in murine and in human neutrophils pre-treated with oxalate for 1 h and activated with Zy/op for 1 h. n = 3 mice per group (representative of two independent experiments) and n = 5 donors (representative of five pooled independent experiments). l ROS production by scramble, PKM2-(∆PKM2) or TPI-(∆TPI) deficient NB4 cells activated with Zy/op for 1 h. n = 3 independent experiments. m Working hypothesis. Error bars are mean ± SEM. p values were determined by two-tailed unpaired Student’s t-test (I, k) or one-way ANOVA followed by Tukey’s post hoc test (b, d–h, l). Clip art provided by Biorender. Source data are provided in the Source data file.

Fig. 6. PKM2-deficiency in myeloid cells impairs host defence against S. aureus infection.

a Schematic protocol of peritonitis infection with S. aureus. b Bacterial load in the blood 18 h after challenge in Pkm2^fl/fl and Pkm2^∆Lyz2 mice. n = 7 and 5 mice, respectively. Representative of two independent experiments. c Bacterial load in the exudate 18 h after challenge in Pkm2^fl/fl and Pkm2^∆Lyz2 mice. n = 7 and 5 mice, respectively. Representative of two independent experiments. d Neutrophil cell counts 18 h after challenge in Pkm2^fl/fl and Pkm2^∆Lyz2 mice. n = 7 and 5 mice, respectively. Representative of two independent experiments. e Mononuclear cell counts 18 h after challenge in Pkm2^fl/fl and Pkm2^∆Lyz2 mice. n = 7 and 5 mice, respectively. Representative of two independent experiments. f–h Concentration of IL-6, CXCL2, and CCL2 in the plasma and exudate 18 h after S. aureus challenge in Pkm2^fl/fl and Pkm2^∆Lyz2 mice. n = 7 and 5 mice, respectively. Representative of two independent experiments. i Schematic protocol of localised cutaneous infection with S. aureus. j Mean of total lesion size (cm²) ± the standard error of the mean (SEM) evaluated for 11 days. n = 10 mice. Representative of two independent experiments. k Representative pictures of the dorsal skin lesions in the back of Pkm2^fl/fl and Pkm2^∆Lyz2 mice evaluated for 11 days. l Bacterial load in the wound two days after challenge in Pkm2^fl/fl and Pkm2^∆Lyz2 mice. n = 9 mice. Representative of two independent experiments. m Myeloperoxidase activity evaluated in the wound two days after S. aureus challenge in Pkm2^fl/fl and Pkm2^∆Lyz2 mice. n = 9 mice. Representative of two independent experiments. n–p Concentration of CXCL1, TNF-α and IL-6 in the wound two days after S. aureus challenge in Pkm2^fl/fl and Pkm2^∆Lyz2 mice. n = 9 mice. Representative of two independent experiments. Error bars are mean ± SEM. p values were determined by one-way ANOVA followed by Tukey’s post hoc test. Clip art provided by Biorender. Source data are provided in the Source data file.