To examine whether primary human macrophages switch to glycolysis after infection with M. avium, we first measured glucose consumption and lactate secretion in spent medium by nuclear magnetic resonance. Figure 1 shows a significant increase in glucose uptake ( Figure 1A ) and lactate secretion ( Figure 1B ) without an effect on glutamine uptake ( Figure 1C ), confirming the activation of glycolysis in presence of M. avium. The glycolytic switch was further confirmed by targeted mass spectrometric analysis of the intracellular concentration of glucose-6-phosphate (G6P) and fructose-6-phosphate (F6P), the first two intermediates of glycolysis, which revealed a significant decrease in both metabolites when compared to the non-infected control ( Figure 1D , white and red). Importantly, macrophages treated with 100 ng/ml of LPS for 24 h display the same characteristics of glycolysis engagement ( Figures 1A–D , blue) as published by others (9), validating our method of measurement. Contrary to Tannahill et al. who reported an increased uptake of glutamine during LPS challenge of mouse macrophages (8), we did not observe an increase of glutamine consumption in LPS-activated human macrophages ( Figure 1C ) which is in accordance with Ko et al. who have reported that human primary macrophages do not rely on glutamine to respond to LPS (13). We next investigated the link between glycolysis and infection by treating infected macrophages with the glycolysis blocker 2-Deoxy-Glucose (2-DG) ( Figure 1E ). Images did not show any signs of cell death induced by the treatment. Quantification of M. avium intracellular growth after 72 hours of infection revealed a 40% increase of fluorescence intensity in cells treated with 2-DG, showing that glycolysis is indeed required to control the intracellular burden ( Figure 1F ).

The glycolytic pathway leads to the formation of cytosolic pyruvate. Pyruvate can then be converted into lactate by the lactate dehydrogenase in a reaction simultaneously regenerating NAD⁺ from NADH, into alanine by the alanine transaminase, or it can enter the mitochondria via the Mitochondrial Pyruvate Carrier (MPC) (14, 15). Quantification by targeted mass spectrometry did not show a significant accumulation in the intracellular level of pyruvate in macrophages infected with M. avium or treated with LPS when compared to untreated controls ( Figure 2A ), confirming that pyruvate is metabolized. Mills et al. have demonstrated that during LPS activation, mouse macrophages switch to aerobic glycolysis while repurposing the tricarboxylic acid (TCA) cycle activity to generate specific immunomodulatory metabolites (9), which implies that a fraction of the pyruvate formed by glycolysis enters mitochondria. We therefore hypothesized that pyruvate transport into mitochondria is the first step of the anti-microbial program centered around mitochondrial metabolism. We investigated the potential contribution of mitochondrial pyruvate to the anti-microbial program by treating infected macrophages with the MPC inhibitor UK5099 (16). Quantification of the images revealed a 68% and 46% increase in M. avium-DsRed signal in cells treated respectively with 100 µM and 10 µM UK5099 compared to the untreated condition ( Figure 2B ), demonstrating that mitochondrial pyruvate is contributing to control M. avium infection in macrophages. We have recently demonstrated that IL-6 and TNF-α, via an auto-paracrine loop, help controlling the intracellular burden of M. avium by supporting the full activation of the Immune Response Factor 1, which leads to the transcription of the mitochondrial enzyme IRG1 and the local production of the metabolite itaconate (7). To test whether glycolysis inhibition or blockade of pyruvate translocation into mitochondria reduces IL-6 and TNF-α production and therefore increases the intracellular burden, we measured their mRNA level by real-time PCR upon 2-DG and UK5099 treatment. We did not observe any alteration of the induction of IL-6 or TNF-α by the infection when compared to the untreated condition ( Figures 2C, D ), on the contrary, we observed a reduction of IL-1β and IL-10 production upon 2-DG treatment, as shown by Tannahill et al. ( Supplementary Figure 1 ) (8). Together, these results demonstrate that pyruvate produced by glycolysis and shuttled in mitochondria contributes directly to the control of the infection.

Once inside the mitochondrial matrix, pyruvate can be converted into acetyl-CoA by the pyruvate dehydrogenase complex (PDC) (17) and feed into the TCA cycle producing reducing equivalents for oxidative phosphorylation. Ruecker et al. have shown that metabolism of fumarate into malate by the mycobacterial enzyme fumarase is required for proper mycobacterial growth but that treatment of M. tb culture with additional fumarate was detrimental (18). Further, we and others have previously found that altered intracellular levels of the TCA cycle-derived metabolite itaconate following an infection was indicative of an anti-microbial function (5, 7) Thus, we decided to probe intracellular concentrations of TCA intermediates to search for indications of a potential antimicrobial metabolite. Targeted mass spectrometric quantification showed stable intracellular levels of all measured TCA intermediates after 24 h of infection, except for a significant reduction in the level of alpha-ketoglutarate (αKG) ( Figure 2E ). We next tested if the infection would affect the energy balance of the cells. Targeted mass spectrometric measurement of the adenine nucleotides AMP, ADP, and ATP, did not reveal any changes after 24 h of infection, leaving the adenylate energy charge, reflecting the energy status of a cell (19), unchanged ( Figure 2F ). To test if the infection level was too low to alter the TCA cycle volume and the energy charge, we used an alternative protocol to virtually infect all the macrophages (MOI 10 for 120 min). Using this condition, we did not observe variation in the intracellular metabolite levels ( Supplementary Figure 2 ). We therefore concluded that none of the TCA intermediates warranted further investigation to explain the anti-microbial effect of glycolysis. Since the infection did not alter the energy charge nor the TCA cycle volume, we hypothesized that pyruvate import, and mitochondrial activity was geared towards mtROS production via the establishment of RET. For this phenomenon to happen, two criteria must be met: the lack of ATP production by oxidative phosphorylation and the maintenance of a high proton motive force/mitochondrial membrane potential (9, 20). To test the engagement of RET, we next treated macrophages infected with M. avium-CFP with the potential-insensitive dye MitoTracker Green and the potential-sensitive dye MitoTracker DeepRed (21–23) ( Figure 2G ). Ratiometric measurement showed that the MitoTracker DeepRed signal was significantly increased by M. avium infection and reduced by the MPC inhibitor UK5099 ( Figure 2H ), suggesting that pyruvate is involved in the establishment and/or maintenance of high proton motive force induced by the infection.

Given that pyruvate import supports the establishment of a high mitochondrial membrane potential and controls M. avium infection, we tested whether pyruvate could be necessary to produce mtROS. Measurement of mtROS production using 500 nM of the mtROS specific dye MitoSOX (24) in the presence of the MPC inhibitor ( Figure 3A ) revealed a significant reduction of the MitoSOX fluorescence intensity in infected macrophages treated with 10 µM UK5099 ( Figure 3B ), indicating that pyruvate shuttling into mitochondria is necessary to produce mtROS. Treatment of infected macrophages with various doses of the mitochondrial specific ROS scavenger MitoTEMPO increased the intracellular burden ( Figure 3C ), confirming that the production of mtROS is an important factor to control the infection. To confirm that mtROS are produced after the engagement of RET, we exploited the ability of the complex I inhibitor Rotenone to reduce the production of mtROS when RET is engaged (25–27). Measurement of mtROS production from macrophages infected with M. avium-CFP and co-treated with 10 nM of Rotenone revealed a 50% decrease of the MitoSOX signal over the non-treated condition ( Figure 3D ), demonstrating both the engagement of RET and the role of complex I during the generation of mtROS. To finally demonstrate the role of RET-related mtROS in the control of the intracellular burden, human macrophages infected with M. avium-DsRed were treated with 10 nM Rotenone for 72 h. Image quantification revealed a 55% increase of the DsRed signal, indicating a higher intracellular burden ( Figure 3E ). The engagement of complex I was confirmed using a high dose of the Carnitine Palmitoyl Transferase 1 inhibitor Etomoxir as an alternative complex I inhibitor (28, 29). Treatment of infected macrophages with 100 µM Etomoxir increased intracellular burden by 52% after 72 h of infection ( Figure 3F ) whereas treatment with the complex II inhibitor Dimethyl Malonate (9) did not significantly affect the DsRed signal ( Figure 3G ). Further, M. avium grown in Middlebrook 7H9 broth in the presence of 2-DG, UK5099, or Rotenone did not show a difference when compared to Middlebrook 7H9 broth alone ( Supplementary Figure 3 ). Finally, we verified that the three different compounds do not interfere directly with the DsRed fluorescence intensity and did not observe any significant variation ( Supplementary Figure 4 ), demonstrating that the effect of these compounds acts through host factors. Collectively, these data suggest that pyruvate import into mitochondria participates to RET-induced mtROS generated by complex I, which contributes to control the intracellular burden.

The nuclear dye Hoechst 33342 was purchased from Life Technologies. Ultrapure LPS (E. coli 0111:B4) was purchased from In vivogen. MitoTEMPO and UK-5099 were purchased from Sigma Aldrich (SML0737; PZ0160). Etomoxir was purchased from Selleckchem (S8244). MitoTracker Green, MitoTracker DeepRed and MitoSOX Red were purchased from Thermo Fisher Scientific (M7514; M22426; M36008).

Buffy coats from healthy blood donors were provided by the Blood Bank, St Olav’s Hospital, Trondheim, after obtaining informed consent and with approval by the Regional Committee for Medical and Health Research Ethics (REC Central, Norway, No. 2009/2245). Peripheral blood mononuclear cells (PBMCs) were isolated using density gradient centrifugation (Lymphoprep, Axis-shield PoC). Monocyte-derived macrophages (MDMs) were generated by plastic adherence for 1h in complete RPMI 1640 (680 µM L-Glutamine and 10 mM Hepes, GIBCO) supplemented with 5% pooled human serum (The Blood Bank, St Olavs hospital) at 37°C and 5% CO₂. After three washing steps with Hank’s Balanced Salt solution (GIBCO), monocytes were cultivated for 6 days with a change of medium at day 3 in complete RPMI 1640/10% human serum and 10 ng/ml recombinant M-CSF (R&D Systems). At day 6 the medium was replaced with complete RPMI 1640/10% human serum and used for experiments on day 7.

M. avium clone 104 expressing CFP or DsRed (used to quantify intracellular burden by measuring fluorescent intensity) was cultured in liquid Middlebrook 7H9 medium (Difco/Becton Dickinson) supplemented with 0.2% glycerol, 0.05% Tween 80 and 10% albumin dextrose catalase. Cultures were maintained at log phase growth (optical density between 0.3 and 0.6 measured at 600 nm, OD600) in a 180-rpm shaking incubator at 37°C for a maximum of 5 days. At the day of infection, bacteria were washed with PBS, sonicated and passed through a Gauge 25 needle to ensure single-cell suspension before challenging day 7 MDMs for 10 min at a multiplicity of infection of 10 (MOI 10). In experiments for Supplementary Figure 2 , MDMs were challenged with M. avium for 120 min. MDMs were subsequently washed with complete RPMI/10% human serum and maintained in culture for the appropriate time. In some experiments, MDMs were challenged with ultrapure LPS (TLR4; 100 ng/ml) for 24 h.

Human MDMs cultivated on glass-bottomed 96 well plates (IBL) were incubated with 10 nM MitoTracker Green and DeepRed for 10 min at 37°C for mitochondria staining or with 500 nM MitoSOX Red for 10 min at 37°C for mtROS staining. MDMs were then washed with complete RPMI/10% human serum and immediately imaged with a confocal microscope.

MDMs cultivated on glass-bottomed 96 well plates were imaged with a Zeiss LSM880 confocal microscope with 40x NA=1.4 objective (Carl Zeiss Micro-imaging Inc.). Emissions were collected using GaAsP hybride detectors. The following acquisition parameters were used: 1024*1024 pixels image size, numerical zoom set to 0.6, frame averaging 1, and 3D acquisition to collect the entire cell with a Z-stack step of 0.25 µm. CFP was excited with a 458 nm Argon laser and emissions were collected through a 470-500 nm window. MitoTracker Green was excited with a 488 nm Argon laser and emissions were collected through a 505-550 nm window. DsRed and MitoSOX Red were excited with a 543 nm HeNe lasers and emissions were collected through a 560-610 nm window. MitoTracker DeepRed was excited with a 633-diode laser and emissions were collected through a 645-700 nm window. Images were analyzed with Image J (NIH).

3D stacks were projected using the “Sum” setting. Resulting images were converted to 8-bit. Regions of interest were drawn around macrophages containing M. avium. The background was estimated using HiLo Lookup Tables and subtracted. For mitochondrial potential, MitoTracker DeepRed values were normalized by MitoTracker Green values. A minimum of 250 infected cells per condition and per donor were counted.

Sampling of MDMs for extraction and mass spectrometric quantification of intracellular central carbon metabolite levels was performed as described for adherent cell lines in (38). In brief, spent medium was discarded and cells were rinsed of extracellular metabolites with cold saline and water on a cold metal block, serving to simultaneously quench metabolism. Next, cells were mechanically detached in cold water:acetonitrile (1:1) and quenched in liquid nitrogen. 5-8 million MDMs were sampled for each donor and treatment. Extraction was performed by cycling between cold water (4°C) and liquid nitrogen, and extracts were concentrated by lyophilization. Absolute quantification of intracellular G6P, F6P, TCA cycle intermediates and adenine nucleotides was performed by capillary ion chromatography tandem mass spectrometry on a Dionex ICS-4000 capillary ion chromatograph (Thermo Scientific) coupled to a Xevo TQ-XS triple quadrupole mass spectrometer (Waters) operating in negative electrospray mode, as described in (39). The modifications described in (40) was applied to allow for isotope dilution and optimized separation of hexose phosphates. Absolute quantification of intracellular pyruvate levels was performed by liquid chromatography tandem mass spectrometry with upfront derivatization, applying an ACQUITY I-Class UPLC coupled to a Xevo TQ-XS triple quadrupole mass spectrometer (Waters) operating in positive electrospray mode, as described in (38). Complete lists of chromatography and mass spectrometry parameters and settings are described in the referenced publications. Data processing was performed in TargetLynx application manager of MassLynx 4.1 (Waters). Absolute quantification was performed by interpolation of calibration curves prepared from serial dilutions of an analytical grade standards calculated by least-squares regression with 1/x weighting. Response factors of all analytical standards and biological extracts were corrected by the corresponding response factor of their U¹³C-labeled isotopologue. Extract concentrations were normalized to seeding density and to the experimental cell volume (3.94 picoliters) measured by a Moxi Z automated cell counter (Orflo) to obtain intracellular concentrations.

Energy charge (EC) was calculated from normalized concentrations of AMP, ADP and ATP using the following formula: EC=(AMP+(0.5*ADP))/(AMP+ADP+ATP) (19).

Extracellular glucose, lactate and glutamine levels of human MDMs were quantified by recording 1D proton NMR spectra of concentrated fresh and spent (24 hours) medium and applying the TopSpin module ERETIC2 (Topspin 4.1.1, Bruker) as described in (41). The difference between fresh and spent medium was normalized to seeding density to obtain consumption and secretion per cell per 24 hours.

MDMs were washed with cold PBS and lysed in buffer RLT (Qiagen) with 1% β-mercaptoethanol. Total RNA was extracted using RNeasy Mini kit according to the manufacturer’s protocol (Qiagen), including DNase I digestion (RNase-free DNase set, Qiagen). The samples included in the study presented an OD_260/280 ratio ~ 2 assessed using a ND-1000 spectrophotometer (NanoDrop). cDNA was synthetized from normalized amounts of RNA using the High-Capacity RNA-to-cDNA kit according to manufacturer’s recommendations (Applied Biosystems). qPCR reactions were performed in 20 µl total volume with 10 ng cDNA input, PerfeCta qPCR FastMix, UNG, ROX (Quanta Biosciences) and TaqMan Gene Expression Assays (Applied Biosystems): GAPDH (Hs99999905_m1), TNF-α (Hs00174128_m1), IL-6 (Hs00985639_m1). The targeted genes were amplified with a StepOnePlus Real-Time PCR System and relative quantities of gene expression were calculated using the comparative C_T method with GAPDH gene expression as endogenous control.

Normality was tested for each experiment. Two-tailed t-test and analysis of variance (ANOVA) were used on normally distributed data; Mann and Whitney test was used otherwise. Significant P values were set as follows: * <0.05, ** <0.01 and *** <0.005. Statistical analyses were performed using GraphPad Prism 8 (GraphPad Software, Inc.).