Neutrophils Fuel Effective Immune Responses through Gluconeogenesis and Glycogenesis

1Neutrophils are the most abundant leukocyte, constituting around 60% of the circulating pool. These short-lived cells are the first to accumulate at inflamed tissues where their primary role is to phagocytose and degrade invading pathogens. The requirement to access ATP rapidly to enable an effective host pathogen response places very specific energy requirements upon these cells. In diseases such as chronic obstructive pulmonary disease (COPD), dysfunctional neutrophilic inflammation is associated with impaired pathogen control and prolonged inflammation, which contribute significantly to pathology. Changes in metabolic flux have been associated with murine macrophage polarization and T cell receptor activation, with associated alteration of immune cell function. How neutrophils dynamically regulate metabolic pathways to fuel antimicrobial responses in the tissues where nutrient supply may be limited remains an important question in the field.

2The existing dogma regarding neutrophil metabolism dates from in vitro studies in the 1950s, detailing the importance of glucose-fueled glycolysis for phagocytosis and respiratory burst activity. Neutrophils cultured in glucose-enriched media utilized extra-cellular glucose for glycolysis and had very low rates of oxidative phosphorylation, in keeping with observations that inhibition of oxidative phosphorylation has little effect on neutrophil oxygen consumption or reactive oxygen species (ROS) production, whereas inhibition of glycolysis using 2-deoxyglucose (2-DG) severely impairs their phagocytic and bacterial killing functions. More recently, studies of human and murine deficiencies in the G6P transporter and G6Pase-β, respectively, have reported that defective intra-cellular glucose cycling can also impact key neutrophil functions. In these settings, impaired glucose shuttling between the endoplasmic reticulum and cytoplasm results in neutropenia and defects in respiratory burst activity, ATP production, and bacterial killing. There is also evidence, however, that a disproportionate increase in glycolysis fuels neutrophil persistence in inflamed airways, resulting in lung injury and death. Neutrophils must therefore tightly regulate glucose utilization and flux through glycolysis to match energy provision to the immediate energy demands of a proportionate but effective host response.

3Neutrophil dependence upon glycolysis allows these cells to function and survive at inflamed sites, where limited oxygen availability may render oxidative metabolism ineffective at meeting energy demand. Since extracellular glucose may also be limited at inflamed sites, we questioned whether neutrophils have a more complex metabolism than previously thought to ensure they can meet energy demand independently of immediate extracellular substrate availability. In this regard, it is interesting to note that neutrophils have been shown to contain glycogen within granules, with early evidence that neutrophil glycogen stores can be modified between circulating and recruited cells, by activation of the oxygen sensing response, and by stimulation with proinflammatory mediators. In this work, we define the metabolic programs that enable neutrophils to generate intra-cellular energy stores and retain function and survival responses under conditions of physiological stress and activation. Further, we demonstrate dysregulation of these metabolic processes in chronic inflammatory disease, with consequences for neutrophil function and survival.

4We used radioactively labeled substrates to characterize which metabolic pathways neutrophils utilize in basal and LPS-stimulated states under conditions of both normoxia and hypoxia (Figures 1A–1C). In keeping with the literature, we observed unstimulated neutrophils to be highly glycolytic, with equivalent glycolytic capacity in normoxia and hypoxia (Figure 1A). LPS stimulation significantly increased glycolytic flux, and this could be reversed by inhibition of glycolysis with 2-DG (Figure 1A). Although detectable levels of glucose oxidation and pentose phosphate pathway (PPP) activity were seen, these were 100-fold less than flux observed through glycolysis (Figure 1B). While at baseline neutrophils had a relatively high level of fatty acid oxidation (FAO), this was suppressed by LPS stimulation (Figure 1C). Metabolic pathways required for ATP and redox are summarized in Figure 1D. To more specifically delineate the effects of physiological stressors on these pathways, we undertook liquid chromatography-mass spectrometry (LC-MS) quantification of individual metabolic intermediaries. We observed that LPS increased relative abundance of key glycolytic metabolites, glucose-6-phosphate/fructose-6-phosphate (Figure 1E), dihydroxyacetone phosphate (DHAP) (Figure 1F), glyceraldehyde-3-phosphate (Figure 1G), and lactate (Figure 1H). In addition, levels of the PPP substrates ribose-5-phosphate (Figure 1I) and sedoheptulose-7-phosphate (Figure 1J), an important branch pathway of glycolysis that enables neutrophil respiratory burst activity, were also induced in response to LPS. In keeping with the increased flux through the PPP and therefore redox buffering capacity, LPS stimulation also significantly increased levels of NADPH and NADH under hypoxic culture (Figures 1K and 1L). Given that neutrophil supplementation with glutamine has previously been described to promote neutrophil phagocytosis, we also analyzed glutamine availability in un-stimulated and stimulated neutrophils under hypoxic and normoxic culture. Although we were unable to detect a significant increase in the total abundance of glutamine in response to LPS (Figure 1M), we found that LPS upregulated glutamine conversion into glutamate (Figure 1N), indicative of increased glutamine utilization in LPS-stimulated neutrophils.

5Radioactive flux assay data were used to predict the relative contribution of FAO, oxidative phosphorylation via the tricarboxylic acid (TCA) cycle, and glycolysis to neutrophil production of ATP. These calculations supported glycolysis contributing a significantly higher proportion of ATP production to neutrophils than FAO under both normoxic and hypoxic culture and following LPS stimulation (Figures 2A and 2B). To validate these calculations, we measured the energy status of neutrophils (ATP/ADP ratio) in the presence of chemical inhibitors of glycolysis (2-DG) (Figure 2C), FAO (etomoxir) (Figure 2D), and oxidative phosphorylation (oligomycin A) (Figure 2E). Only treatment of neutrophils with the hexokinase inhibitor 2-DG resulted in a failure of neutrophils to maintain their high energy status, independent of oxygen availability (Figures 2C–2E). It is important to note that 2-DG may be exerting off-target effects through inhibition of N-glycosylation and induction of the unfolded protein response. To assess whether neutrophils have the capacity to actively maintain favorable energetics, we measured relative levels of ATP, ADP, and AMP over time, and determined the energy charge ([ATP + 1/2ADP]/[ATP + ADP + AMP]) of neutrophils under conditions of hypoxia, glucose deprivation, and LPS stimulation. Neutrophils were able to maintain their energetic balance across multiple time points (Figure 2F) even when oxygen and glucose were limited (Figures 2G and S1A–S1C). Neutrophils are therefore adept at maintaining intracellular energy homeostasis both in quiescent and activated states, throughout their lifespan, and despite the presence of physiological stresses including glucose and oxygen deprivation, provided they can maintain an active glycolytic pathway.

6Having demonstrated neutrophil reliance on glycolysis, even in the absence of extra-cellular glucose, we next questioned whether neutrophils can meet their energy demands by generating and utilizing intracellular glucose stores in the form of glycogen. Glycogen levels in unstimulated and LPS-stimulated cells were measured under conditions of normoxia and hypoxia. LPS-stimulated neutrophils had significantly higher levels of glycogen in both normoxia and hypoxia (Figure 3A). Inhibiting glycogen degradation using the glycogen phosphorylase inhibitor CP-91149 increased intracellular glycogen content (Figure 3B), and the resulting inability of neutrophils to access glycogen stores led to impaired neutrophil survival in glucose-deprived cultures of both non-LPS- and LPS-treated cells (Figure 3C).

7Combining the ability of neutrophils to fuel glycolysis and maintain energy states in glucose-deplete conditions with their ability to regulate intracellular glycogen led us to question whether neutrophils can utilize glucose and non-glucose substrates to generate glycogen stores. We first verified that neutrophils regulate glycogenesis by demonstrating increased uptake of U-14C glucose and higher levels of radioactively labeled glycogen following LPS challenge (Figure 3D). Importantly, when neutrophils were cultured with U-14C lactate, we were again able to detect significant levels of radioactivity in the purified glycogen (Figure 3E), indicating that neutrophils can use non-glucose substrates to generate glycogen stores. In keeping with the ability of neutrophils to undergo glycogenesis and gluconeogenesis, we observed that neutrophils express key enzymes required for active glycogenesis (GYS1, GBE1, and UGP2) and gluconeogenesis (FBP1 and PEPCK2) at both mRNA and protein levels (Figures 3F and 3G).

8To confirm that neutrophils have the capacity for gluconeogenesis, cells were incubated with heavy U-13C glucose, U-13C glutamine, and U-13C palmitate tracers, and analysis of 13C labeling was quantified by LC-MS. A schematic detailing the putative entry and cycling of carbon molecules from U-13C glucose (black circles) and U-13C glutamine (gray circles) into TCA and glycolytic isotopologues is provided in Figure 3H. Detection of multiple isotopologues of glucose-6-phosphate/fructose-6-phosphate following neutrophil culture with U-13C glucose (Figure 3I) and the presence of heavy carbons within glycolytic and TCA intermediaries following neutrophil culture with U-13C glutamine (Figures 3J–3L) and U-13C palmitate (Figure 3M) together provide definitive evidence for active neutrophil gluconeogenesis. The ability of neutrophils to induce gluconeogenic activity was observed with both LPS treatment (Figures 3I–3L) and in hypoxia (Figures 3J–3L), with higher levels of labeled carbons from glutamine observed in glycolytic intermediaries (Figure 3J). Thus, glutamine contributes to the increase in gluconeogenesis observed in response to hypoxia and LPS. The full isotopologue profiles for citrate/isocitrate, oxaloacetate, malate, pyruvate, and F1,6BP following neutrophil culture for 4 h with U-13C glutamine are shown in Figures S2A–S2F, with significant differences in substrate utilization observed within 2 h of U-13C glutamine supplementation (Figures S3A–S3F). Increased tracing of U-13C glutamine carbons into PPP metabolites was also observed with LPS stimulation (Figures S4A and S4B). In order to directly assess whether gluconeogenesis contributes to the synthesis of glycogen, we undertook the tracing of heavy carbons from pyruvate into glycogen. 13C-labeled glucose generated from hydrolyzed glycogen was observed in LPS-stimulated neutrophils (Figure 3N), providing direct evidence for a role of gluconeogenesis in fueling glycogen synthesis.

9With evidence for dynamic regulation of both gluconeogenesis and glycogenesis, we questioned the interplay between these pathways and their importance for neutrophil function and survival. Culture of LPS-stimulated neutrophils with the glutaminase inhibitor BPTES significantly reduced intracellular glycogen levels, supporting a role for glutaminolysis in the synthesis of glycogen (Figure 4A). Furthermore, neutrophil intracellular glycogen levels were also significantly reduced following inhibition of gluconeogenesis (FBP1) with the inhibitor MB05032 (Figure 4B). The functional consequence of limiting glutamine access and blocking gluconeogenesis was next explored. Glutamine deprivation and inhibition of either PEPCK (3-mercaptopicolinic acid, 3-MPA) or FBP1 (FBPi or MB05032) resulted in reduced neutrophil survival (Figures 4C–4E). Neutrophils also displayed defective bacterial killing when gluconeogenesis (Figure 4F) and glutaminolysis (Figure 4G) were disrupted. To directly test whether gluconeogenesis and glycogen shuttling regulate neutrophil function in vivo, we generated a neutrophil-specific glycogen synthase (Gys1) knockout mouse (Figures S5A and S5B). Targeting GYS1, the dominant isoform in neutrophils, enabled us to prevent the generation of glycogen in neutrophils, therefore abrogating the glycogen shuttle. Gys1loxloxMRP8-cre+/− mice demonstrated no baseline defect in blood neutrophil numbers (Figure S5C) or monocyte proportion (Figure S5D). Airspace neutrophils were isolated from mice 24 h post-LPS challenge, and the consequences of GYS1 loss for apoptosis and bacterial killing were assessed ex vivo. In order to identify the neutrophil niche in which glycogen is most important, glycogen was first measured in bone marrow (BM), blood, and bronchoalveolar lavage (BAL) neutrophils. We found that glycogen levels in the BM and circulating neutrophils of naive and LPS-challenged wild-type (WT) mice are significantly lower than those of BAL neutrophils recruited following LPS challenge (Figure 4H). Subsequent assessment of glycogen content in blood and airspace neutrophils from Gys1loxloxMRP8-cre+/− and Gys1loxloxMRP8-cre−/− mice 24 h post-LPS challenge confirmed that Gys1 loss results in a significant reduction in levels of glycogen in these cells (Figure 4I). In vitro, this translated into functional consequences shown by an increase in apoptosis (Figure 4J) and reduced bacterial killing (Figure 4K). Importantly, Gys1loxloxMRP8-cre+/− mice revealed an impaired host response following in vivo challenge with subcutaneous administration of Staphylococcus aureus (Figures 4L–4N). Systemically, these mice developed a significant degree of hypothermia (Figure 4L), with increased bacterial counts within the skin abscess 24 h following bacterial challenge (Figure 4M), and reduced circulating neutrophil and monocyte numbers (Figures 4N and S5E).

10With evidence that glycogen shuttling is important for neutrophil-mediated host responses, we questioned whether changes in glycogen stores had long-term consequences for neutrophil survival. Recent work from our group has shown that sustained exposure of mice to systemic hypoxia reprograms neutrophil glucose utilization, and that activation of oxygen sensing responses can alter neutrophil glycogen stores. We therefore hypothesized that exposure of healthy human volunteers to altitude-induced hypoxia would enhance the capacity for activated neutrophils to generate glycogen stores with long-term consequences for neutrophil survival. Twelve healthy human volunteers were exposed to 7 days of altitude-induced hypoxia (Figures S5F and S5G) and their peripheral blood neutrophils were isolated at baseline (BL), at altitude, and 3 months post altitude (PA), then cultured in the presence or absence of LPS. Altitude exposure resulted in significantly increased neutrophil glycogen stores following LPS stimulation (Figure 4O), driven by induction of UGP2 and glycogen synthase GYS1 (Figures 4P, 4Q, S5H, and S5I). This inflammatory uplift in glycogen stores following altitude exposure was paralleled by enhanced neutrophil survival responses when compared to baseline (Figures 4R, S5K, and S5L).

11Disordered neutrophilic inflammation is a characteristic feature of COPD, with bacterial persistence and ongoing neutrophil recruitment associated with more rapid disease progression. A disconnect between high neutrophil numbers and failed pathogen clearance in this setting is indicative of aberrant neutrophil function, a phenotype of considerable therapeutic interest. Peripheral blood neutrophils from patients with GOLD stage 1–4 COPD (Table S1) and no evidence of systemic hypoxia and respiratory failure were studied. Patient neutrophils had impaired bacterial killing of Streptococcus pneumoniae S14 (Figure 5A) and S. aureus SH1000 (Figure 5B), with defective LPS survival responses (Figure 5C). Importantly, impaired survival responses could not be rescued by glucose supplementation, indicating an intrinsic neutrophil defect independent of glucose availability. In keeping with a dysregulated metabolic status, COPD neutrophils had significantly lower ATP levels at baseline when compared to healthy controls (Figure 5D) and a reduced glycolytic response to SH1000 (Figures 5E and 5F). Diminished glycolytic capacity occurred despite preservation of glucose uptake with equivalent tracing of U-13C glucose into glucose-6-phosphate (Figure 5G) and retained expression of the dominant neutrophil glucose transporter GLUT1 (Figure S6A) in unstimulated and LPS-stimulated states. Extending our observations that neutrophils have an active gluconeogenesis pathway that uses non-glucose substrates to generate glycogen, and the importance of glycogen stores for neutrophil survival, we therefore questioned whether defective gluconeogenesis and glycogen cycling could result in these dysfunctional neutrophil responses. Initial baseline phenotyping of peripheral blood neutrophils revealed an impaired ability of COPD neutrophils to increase glycogen levels in response to LPS (Figure 5H) and to increase transcription of the key gluconeogenic enzyme PEPCK2 (Figure 5I) and the glycogenic enzyme GBE1 (Figure 5J). LC-MS analysis of healthy control and COPD blood neutrophils was then undertaken to delineate the metabolic consequences of impaired glycogen cycling in these cells. COPD neutrophils displayed significantly lower lactate levels with normoxic culture alone and following LPS stimulation (Figure 5K). This was associated with a decreased level of lactate in the medium of COPD neutrophils cultured with LPS (Figure 5L) and reduced levels of the PPP substrates ribose-5-phosphate (Figure 5M) and sedoheptulose-7-phosphate (Figure 5N). COPD neutrophils also demonstrated a failure to generate the metabolites required for effective redox buffering capacity (NADH) (Figure 5O) and pathogen sensing responses (UDP-glucose and UDP-GlcNAc) (Figures 5P and 5Q). Finally, tracing of U-13C glutamine into F1,6BP (Figure 5R) demonstrated that COPD neutrophils have a diminished capacity to utilize glutamine for gluconeogenesis.

12Figure 6 summarizes the dynamic balance between neutrophil energy utilization and storage driven by intracellular glycogen stores in quiescent and activated blood neutrophils, and how this becomes uncoupled in chronic inflammatory states.

13Cellular adaptation to physiological stresses encountered within inflamed tissues is critically dependent upon metabolic responses that maintain ATP and modify energy requirements. Neutrophils are unusual in that they must rapidly access ATP to enable responses, including migration, pathogen control, and apoptosis, that have significant energy requirements. As such, they have evolved a dependence upon glycolysis for ATP production. We have delineated the metabolic processes that enable this metabolic specialization, allowing successful matching of energy supply and demand through the ability of neutrophils to regulate autocrine energy stores in the form of glycogen. Through cycling between gluconeogenesis and glycogenesis, we propose that neutrophils are able to maintain their energy charge throughout their lifespan, even when glucose is limited.

14Initial experiments detailing the existence of multiple glucose isotopologues with U-13C glucose tracing, and demonstrating neutrophil expression of unidirectional enzymes required for gluconeogenesis (PCK2 and FBP1), led us to propose that neutrophils can undergo gluconeogenesis. This was verified using LC-MS tracing of heavy carbons from U-13C glutamine and palmitic acid into glycolytic intermediaries and finally U-13C pyruvate into glycogen. The ability of neutrophils to scavenge glutamine for generation of glycolytic intermediaries is particularly important, given the relative availability of these metabolic substrates at the inflamed site, as this provides a mechanism whereby neutrophils can continue to meet energetic demands in the absence of extracellular glucose. The capacity for limited non-hepatic gluconeogenesis is not entirely without precedent. In a glucose-deprived setting, cancer cells have the ability to switch to using glutamine as a substrate and engage the early steps of gluconeogenesis through regulation of mitochondrial PCK2. More recently, maintenance of CD8 T cell memory has been linked to activation of the cytosolic form of this enzyme PCK1, while loss of PCK1 has also been associated with a reduction in TCA cycle intermediaries and pro-inflammatory macrophage polarization. There is, however, no evidence to date suggesting that myeloid populations other than neutrophils have the capacity to undergo gluconeogenesis and only a single report demonstrating the differential regulation of glycogen and glucose metabolism in dendritic cells, indicating cell-type-specific metabolic adaptations within the innate immune response and thus the opportunity for neutrophil-specific therapeutic intervention.

15We subsequently questioned how neutrophils balance energy utilization with energy storage and the consequences for immune cell function. This was explored by exposing peripheral blood neutrophils from healthy volunteers to differing physiological stresses (glucose depletion and hypoxia), pro-inflammatory activation with LPS, and inhibitors of glycogen degradation. Increased generation of glutamine-derived glycolytic intermediaries together with increased gluconeogenesis fueled glycogen deposition following LPS stimulation of neutrophils would suggest that in response to a proinflammatory stimulus neutrophils increase cycling through gluconeogenesis and glycogenesis to maintain glycogen storage, which in turn fuels the higher energy demands of a proinflammatory response.

16Importantly, this glycogen/glucose axis is both amenable to modulation and regulates neutrophil function. Exposing healthy volunteers to altitude-induced hypoxia resulted in augmented intracellular glycogen reserves in response to LPS and enhanced neutrophil survival. The ability of cells to shunt glucose into glycogen has previously been observed in brain astrocytes and cancer cells. In hepatoma cell lines, these changes in glycogen turnover are sensitive to hypoxia and important for cell survival. More broadly, glycogen utilization has also been linked to cancer cell proliferation, abrogation of premature senescence, and optimal functioning of the PPP. The impact of hypoxia in this setting is of considerable interest given hypoxia is an important hallmark of the tumor microenvironment. This is further strengthened by the observation that glycogen biosynthetic enzymes (GYS1, GBE1, and UGP2) are themselves regulated at a transcriptional level by hypoxia. In a disease setting, defective glutamine utilization for gluconeogenesis and reduced glycogen shuttling result in insufficient glycogen stores to fuel effective immune responses and generate pathogen-sensing sugars, with consequences for neutrophil survival and bacterial killing. Moreover, in vivo study of neutrophils deficient in the glycogen synthesis enzyme GYS1 provides direct evidence that neutrophil glycogen fuels effective host immunity, with impaired bacterial control observed in Gys1loxlox MRP8-Cre+/− transgenic mice.

17Our study demonstrates that neutrophils have adapted to use multiple metabolic substrates to generate energy stores in the form of glycogen, which are dynamically regulated by both gluconeogenesis and glycogenesis. Increased glycogen shunting enables neutrophils to meet the higher metabolic demands of a pro-inflammatory response. Conversely, failure to replenish or access these glycogen stores results in metabolic exhaustion and dysfunctional neutrophil responses. What underpins the defective glycogen cycling we observe in neutrophils isolated from patients with chronic inflammatory disease is currently unknown, and a limitation of the work undertaken to date. Moreover, how neutrophil metabolic responses are reprogrammed and retained within circulating populations over time requires further study. Nonetheless, we propose that this metabolic specialization of neutrophils could allow selective targeting of glycogen cycling to optimize the efficacy of neutrophil host responses and thus facilitate inflammation resolution.

18Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, sarah.walmsley@ed.ac.uk

19The mouse line (Gys1lox/lox MRP8-Cre+/−) generated in this study is available from the Lead Contact upon request.

20This study did not generate/analyze datasets/code.

21Peripheral venous blood neutrophils were isolated from healthy volunteers with written informed consent as approved by the CIR Blood Resource Management Committee (AMREC 15-HV-013) in Edinburgh and the Commisie Medische Ethiek UZ KU Leuven in Belgium (total number of participants 67 aged between 20-70 years, mean age range 31-40, Gender M 27: F 40). Peripheral venous blood neutrophils were also isolated from patients with COPD in accordance with local ethics (REC reference: 15/SS/0095). Patient details are shown in Table S1 (total number of participants 24 aged 40-76 years, mean age 58.7, Gender M 10: F 14). Exclusion criteria included history of anemia, pregnancy or breast-feeding, Diabetes Mellitus, CKD, hepatic failure, immunosuppression and febrile illness within the last two weeks. COPD patients were matched to healthy non-smoker control donors ± 5 years of age wherever possible (total number of participants 18 aged between 20-70, mean age range 41-50, Gender M 8: F 9). Negative magnetic selection (STEMCELL Technologies, Vancouver) was carried out to further purify the neutrophils and consistently demonstrated purity of over 98%.

2212 healthy participants aged 19-22 years (mean age 20, Gender M 6:F 6) were recruited from the University of Edinburgh student community. Exclusion criteria included previous hospital admission for asthma, significant cardio-respiratory disease, regular cardiovascular medications, pregnancy, smoking and previous exposure to high altitude (defined as > 2500 m). Ethical approval was granted by the ACCORD Medical Research Ethics Committee (REC reference: 17-HV-030) and all participants provided informed written consent. Data were collected at sea level, during baseline and post-expedition testing, and at 4,700 m on the APEX 5 expedition. Subjects flew into La Paz, Bolivia (3,600 m) within a 12 h period and after acclimatizing for four days, ascended 90 min by bus to Huayna Potosí base camp (4,700 m) for seven days before returning to La Paz. Mean peripheral blood oxygen saturation (SpO2) was measured daily using pulse oximetry (Santa Medical, Tustin). In all control and patient cohorts both male and female participants were included and matched where possible, small sample sizes precluded any further study of the influence of sex.

23Neutrophils were isolated from sodium citrate anticoagulated blood as previously described using dextran sedimentation and discontinuous percoll gradients.

24Gys1 lox/lox mice were crossed with the previously reported neutrophil specific cre driver Mrp8 to generate Gys1lox/lox MRP8-Cre+/− mice on a C57BL/6JOla genetic background. Both male and female littermates were used from ages 6-12 weeks old with Gys1lox/lox MRP8-Cre−/− mice and wildtype C57BL/6JOla sex matched or litter mate controls.

25Animal experiments were conducted in accordance with the UK Home Office Animals (Scientific Procedures) Act of 1986. All animal studies were approved by The University of Edinburgh Animal Welfare and Ethical Review Board. 6-12 week old male and female mice were used for experiments. Mice were housed in IVC cages under 12 h light/darkness cycles and controlled temperature (20-23°C) in accordance with UK Home Office guidance. All mice used for experiments were healthy with quarterly and annual testing carried out in accordance with FELASA 2014 Guidelines, using a mixture of environmental, random colony samples and sentinel testing by serology and PCR. Mice had free access to food (Special diets service rat and mouse number 1 maintenance food RMI (P) 801151) and water.

26Neutrophils were cultured at 5 million cells/mL for 2-20 h at 37°C, in normoxia (21% O2, 5% CO2) or hypoxia (1% (3kPa) O2, 5% CO2) in two types of pre-equilibrated RPMI media 1640 with 10% fetal bovine serum (FBS) (GIBCO) and 1% Penicillin/Streptomycin, differing in the presence or absence of glucose. In the absence of glucose, dialyzed glucose-free FBS was used. To study the effects of stimulation and metabolic inhibitors, neutrophils were cultured with (100ng/mL) LPS E.coli serotype R515 (Enzo), CP-91149 (Sigma-Aldrich), 2-deoxyglucose (Sigma-Aldrich), Oligomycin A (Sigma-Aldrich) and Etomoxir (10uM) (CNIO Carlos III Therapies), Fructose-1,6-bisphosphatase-1 Inhibitor (FBP1i) (Cambridge Bioscience), MB05032 (Tocris Bioscience), 3-Mercaptopicolinic acid (HCl) (3-MPA) (Cambridge Bioscience), Glutaminase Inhibitor II, BPTES (MERCK Chemical).

27Purified peripheral blood neutrophils (2.5 x106) were incubated for 6 h in RPMI 1640 medium (supplemented with 5.5 mM unlabeled glucose, 10% FCS and 1% Penicillin/Streptomycin) containing 0.4 μCi/mL [5-3H]-D-glucose (Perkin Elmer). Cells were pelleted (420 g for 10 min) and supernatant transferred into glass vials containing 12% perchloric acid sealed with rubber stoppers. 3H2O was captured in hanging wells containing a piece of Whatman paper soaked with H2O over a period of 48 h at 37°C to reach saturation. Radioactivity in the Whatman paper was determined by liquid scintillation counting.

282.5 x106 were incubated 6 h in RPMI 1640 medium (supplemented with 5.5 mM unlabeled glucose, 10% FCS and 1% Penicillin/Streptomycin) containing 0.4 μCi/mL [-3H]-D-palmitic acid (Perkin Elmer).

29The contribution of the TCA cycle to total glucose utilization was measured as 14CO2 release formation both in the TCA cycle and oxidative pentose phosphate pathway using 0.55 μCi/mL [6- 14C]-D-glucose. Following a 6 h incubation in RPMI 1640 medium (supplemented with 5.5 mM unlabeled glucose, 10% FCS and 1% Penicillin/Streptomycin) containing 0.55 μCi/mL [6- 14C]-D-glucose, 2.5 x106 cells were lysed using 250 μL of 12% Perchloric acid. Wells were immediately covered with a piece of 1x hyamine hydroxide-saturated Whatman paper and overnight absorption of 14CO2 released was performed at room temperature. Radioactivity in the Whatman paper was measured using scintillation counting.

30A total of 5x106 purified human peripheral blood neutrophils were harvested in 100 μL of ice-cold 5% PCA supplemented with 1 mM EDTA. ATP, ADP and AMP levels were measured using ion-pair RP-HLPC. Thereafter, samples were centrifuged for 10 min at 13000 g and the supernatant neutralized with K2CO3. Following perchlorate precipitation, the supernatant was injected on a 4.6 × 150-mm, 5-μm particle size C-18 HPLC column at a rate of 1 mL/min, 100% buffer A from 0 to 5 min, 100% buffer A to 100% buffer B from 5 to 20 min, 100% buffer B from 20 to 31 min for column re-equilibration (buffer A: 25 mM NaH2PO4, 0.385 mM tetrabutylammonium, pH 5; buffer B: 10% (v/v) acetonitrile in 200 mM NaH2PO4, 0.385 mM tetrabutylammonium, pH 4). Phosphorylated nucleotides were monitored at 260 nm. The energy charge was expressed as ([ATP] + 1/2 [ADP] / [ATP] + [ADP] +[AMP]) and the energy status of the cells as the ratio of ATP to ADP content.

31Ultrapure neutrophils were cultured in the presence of 5.5 mM. U-13C glucose or 100 μM U-13C palmitate for 4 h and 2mM U-13C glutamine for 2,4 and 6 h. 2 0.5x106 cells were harvested in 100 μL of 80% methanol. Measurements of relative levels of analyte abundance and 13C incorporation into glycolytic intermediates were performed using a Dionex UltiMate 3000 LC System (Thermo Scientific) coupled to a Q Exactive Orbitrap Mass Spectrometer (Thermo Scientific) operated in negative mode. Practically, 25 μL of sample was injected on a SeQuant ZIC/pHILIC Polymeric column (Merck Millipore). The gradient started with 10% solvent B (10 mM NH4-acetate in mqH2O, pH 9.3) and 90% solvent A (acetonitrile) and remained at 10% B until 2 min after injection. Next, a linear gradient to 80% B was carried out until 29 min. At 38 min, the gradient returned to 40% B, followed by a decrease to 10% B at 42 min. The chromatography was stopped at 58 min. The flow was kept constant at 100 μl/min, and the column was kept at 25°C throughout the analysis. The MS operated in full scan–SIM mode using a spray voltage of 3.2 kV, capillary temperature of 320°C, sheath gas at 10.0, auxiliary gas at 5.0. AGC target was set at 1e6 using a resolution of 140,000, with a maximum IT of 500 ms. Data collection and analysis were performed using Xcalibur Software (Thermo Scientific). Total metabolite abundance is shown as a measure of both unlabeled 12C isotopologues and all other mass isotopologues corresponding to the analyte of interest. Isotope correction was carried out as previously described using an in-house software tool.

32Neutrophils (10 x106) were cultured in U-13C pyruvate for 8 h under normoxia and hypoxia. The cells were washed with saline and lysed by boiling for 15 min in 30% KOH. Glycogen was precipitated using ice cold 100% ethanol and washed a further two times. Following the evaporation of residual ethanol and resuspension in water the glycogen was hydrolysed using amyloglucosidase (Merck Life Sciences UK). Acetonitrile:methanol:water 40:40:20 was added to the samples and samples analyzed by LC-MS.

33Neutrophils (5 × 106) were culture with U-14C glucose or U-14C lactate for 6 h in the absence and presence of LPS under normoxia and hypoxia. The cells were washed 5 times in ice-cold PBS and lysed in 30% KOH by boiling for 15 min. An aliquot of the homogenate was spotted onto Whatman paper and the glycogen precipitated by immersing the paper in ice-cold 66% ethanol (v/v). Following two washes with ethanol, radioactivity in the Whatman paper was measured using scintillation counting.

341 × 106 human neutrophils were lysed in 200 μL ice-cold H2O and boiled for 10 min at 95 degrees; lysates were centrifuged at 18,000 g at 4°C for 10 min to remove cell debris and snap frozen. Glycogen content was measure using a fluorimetric assay (Sigma-Aldrich).

35Neutrophils were resuspended in XF assay media at a concentration of 3x106/mL onto a XF24 cell plate pre-coated with Cell-Tak (Corning). Cells were stimulated with Staphylococcus aureus (SH1000) at MOI 10, 25 and 50. The ECAR were measured at intervals of 7 min over a 90-min cycle using a Seahorse XF24 (Seahorse Bioscience USA).

36Neutrophil apoptosis following 20 h of culture was measured by flow cytometry of cell stained with Annexin V PE and the nuclear dye TO-PRO-3 (Molecular Probes, Thermo Fisher). Flow cytometric measurements were performed using a four-color FACSCalibur flow cytometer (Becton Dickinson).

37RNA was isolated from neutrophils using the mirVana total RNA isolation protocol (Ambion, Thermo Fisher), DNase treated and reverse transcribed using AMV reverse transcriptase with random primers (Promega). TaqMan gene expression assays (Applied Biosystems, Thermo Fisher) and PrimeTime qPCR Probe Assays (IDT) were used for relative quantification of cDNA using SDS 2.4 (Thermo Fisher) and normalized to ACTB expression.

38Neutrophils were infected for 90 min with an MOI of 10. After the killing of extraceullar bacteria for 30 min with either Gentamycin (20 μg/mL) and Benzylpenicillin (40 units/mL) when using serotype 14 Streptococcus pneumoniae (S14) or Gentamycin (20 μg/mL) and Vancomycin (10 μM) when using Staphylococcus aureus (SH1000), the numbers of internalised bacteria at times 0 and 3 h were measured.

39Neutrophils were cultured for 4 h in glucose-deprived media with the inhibitors and infected with Staphylococcus aureus SH1000 for 60 min with an MOI of 10. After the killing of extraceullar bacteria for 30 min with either Gentamycin (20 μg/mL) and Vancomycin (10 μM), the number of internalised bacteria at times 0 and 2 h were measured.

40Neutrophil whole cell lysates were prepared with Laemmli buffer containing PI cocktail (protease inhibitor), PhosSTOP, PMSF (phenylmethylsulfonylfluoride) and DFP (Diisopropyl fluorophosphate). Following lysis samples were boiled with 2x loading buffer containing SDS. Immunoblotting was performed with antibodies against glycogen synthase (total), phosphoglycogen synthase (Ser641) (Cell Signaling), glycogen phosphorylase (liver form) (Abcam), phosphoenol pyruvate carboxykinase 2 (PCK2) (Abcam), fructose 1,6 bisphosphatase 1 (FBP1) (Abcam) and HRP-conjugated secondaries: anti-mouse IgG (Cell Signaling Technology) and goat anti-rabbit IgG (Dako). Total P38 was used as a loading control.

41Mice were injected subcutaneously with live stationary phase Staphylococcus aureus (SH1000) at 5x107 CFU. At 24 h, rectal temperatures and abscess areas were recorded, mice anesthetized and exsanguinated, and tissue processed for CFU counts.

42Data represent mean ± SEM from at least three independent experiments.

43Significance was determined by paired 2-tailed t tests unless otherwise stated. All statistical tests were performed with Prism 7 software (GraphPad Software) using unpaired t tests, paired t tests, one-way or two-way ANOVA with Tukey’s multiple comparisons test, two-way ANOVA with Sidak’s multiple comparions test. Statistical parametric analyses were performed following confirmation of normal distribution of the data. ∗p < 0.05 was considered as statistically significant. ∗∗p < 0.01, ∗∗∗p < 0.005, ∗∗∗∗p < 0.001. Statistical parameters are highlighted by ∗ values in each figure with details provided in the associated figure legends.