Loss of mTORC2 Activity in Neutrophils Impairs Fusion of Granules and Affects Cellular Metabolism Favoring Increased Bacterial Burden in Sepsis

Sepsis is a life-threatening syndrome characterized as an uncontrolled inflammatory response to infection, causing injury to tissues and organs (1). The innate immune system has a crucial role in sepsis pathophysiology. Neutrophils have a key role in microbial infections because of their quick response and ability to eliminate the pathogen through their effector functions (2). The full activation of neutrophils occurs after they migrate to the infectious site and encounter the pathogen. Then the neutrophils engulf the pathogen within seconds, thus forming the intracytoplasmic phagosome. At this moment, the granules’ traffic and fusion are essential to generate the phagolysosome (3). Through phagolysosome formation, neutrophils can produce oxidants and eliminate microorganisms. The oxidative burst allows neutrophils to generate potent microbicide agents, such as hypochlorous acid (HOCl) and NO-derived oxidants (4). The step-limiting enzymes in these processes, myeloperoxidase (MPO) and inducible NO synthase, are found in primary azurophilic granules (5). Oxidants are also essential for the release of chromatin web-like structures, called neutrophil extracellular traps (NETs) (6, 7). NETs are important for extracellular pathogen elimination by releasing microbicidal enzymes, such as MPO, neutrophil elastase, and histones, but they also amplify inflammation and recruit other cells (8).

Neutrophils can control their metabolic processes to adapt and optimize their effector functions according to their needs and the nutrients available in the environment (9, 10). Although the main source of energy in neutrophils is aerobic glycolysis, mitochondria form a complex network in the cytoplasm and are vital for the effector functions of these cells, such as chemotaxis and oxidant production (11, 12). Over the last few years, the well-studied mammalian target of rapamycin (mTOR), one of the master regulators of the bioenergetic status of immune cells, has been explored in neutrophils (13). The mTOR pathway comprises two complexes: mTOR complex 1 (mTORC1), which includes the associated regulatory protein of mTOR (Raptor), and its activation is inhibited by rapamycin; and mTOR complex 2 (mTORC2), which involves the rapamycin-insensitive companion of mTOR (Rictor). Due to the rapamycin treatment, the activation, inhibition, and function of mTORC1 are investigated in a variety of immune cells (13–15). By contrast, there is no drug to specifically inhibit mTORC2, so its activation and function remain unclear, especially in neutrophils. Recent studies have demonstrated that mTORC2 acts in the regulation of neutrophil chemotaxis. Activation of mTORC2 is crucial for the cytoskeleton rearrangement via F-actin and Myosin II (16, 17). The mTORC2, in addition to being involved in the actin polarization, is also important to limit actin assembly by a mechanosensory biochemical negative feedback cascade (18). The mTOR pathway can also contribute to neutrophil chemotaxis by controlling the mitochondria ATP production and purinergic signaling (11).

Although other studies have suggested that the axis PI3K/Akt/mTOR is involved in neutrophil effector functions (19, 20), little is known about the role of mTORC2 in these cells, particularly during sepsis onset.

In this study, we hypothesized that the inhibition of the mTORC2/Rictor axis in neutrophils impairs, besides their migration, other key effector mechanisms, thus leading to a poor outcome in a mouse model of sepsis. We found that the absence of mTORC2 signaling led to an improper fusion of granules and HOCl production in neutrophils. We also observed metabolic alterations, such as increased aerobic glycolysis in Escherichia coli–stimulated neutrophils, leading to microbicidal capacity impairment and consequently increased bacterial burden during sepsis. In addition, our bioinformatics analysis of neutrophils isolated from patients with sepsis demonstrated that Rictor is downregulated in these cells and correlates negatively with genes related to the glycolytic pathway. Taking together, our results highlight the importance of a proper mTORC2 activation for neutrophil effector function during sepsis.

All animal experiments were conducted in agreement with federal guidelines and approved by the institutional committee for animal care and use at the University of São Paulo (USP), Institute of Biomedical Sciences, São Paulo, Brazil. This study was registered under the protocol numbers 51/2014 and CEUA.053.2018. Studies were performed on age- and sex-matched littermates. Male mice, 8–12 wk old, were used for in vivo experiments. Female mice, 8–12 wk old, were used for in vitro experiments.

Rictor^flox/flox and LysM-Cre mice were all on the C57BL/6 background. LysM-cre^+/+ transgenic mice [B6.129P2-Lyz2tm1(cre)Ifo/J; The Jackson Laboratory] were kindly provided by Prof. Dr. William Festuccia from the Physiology Department at the USP. Rictor^flox/flox (Rictor^tm1.2Mgn) mice were purchased from The Jackson Laboratory. We generated homozygous Rictor^flox/flox LysM-Cre^+/+ (LysMRic Δ/Δ) mice or Rictor^flox/flox LysM-Cre^−/− (LysMRic fl/fl) mice by crossing Rictor^flox/flox with LysM-Cre. We used the homozygous LysMCre to provide an optimal mTORC2 degree of inactivation in neutrophils, as seen in other disease models using LysMCre mice (21). Animals were maintained in microisolator cages on a rodent ad libitum sterilized chow diet and water. For all mice used, the genotypes were determined by PCR analysis of tail genomic DNA. Rictor deletion was verified through the removal of exon 3 (Rictor gene) by conventional PCR, using genomic DNA from purified bone marrow neutrophils (Supplemental Fig. 1A, 1B) and by the phosphorylation of Akt at serine 473 (mTORC2 site of phosphorylation) by Western blot (Fig. 1A, 1B).

For the in vitro experiments, we used isolated neutrophils from bone marrow. First, we harvested the femurs and tibias from mice and flushed the cells with 5 ml of HBSS (Sigma-Aldrich, St. Louis, MO). After centrifugation (300 × g for 10 min at 4°C), the RBCs were lysed with a hypotonic lysis buffer solution (155 mM NH₄Cl, 12 mM NaHCO₃, 0.1 mM EDTA–EDTA). After neutralizing the lysis buffer, cells were filtered through a 40-μm cell strainer. Subsequently, we purified the neutrophils using the positive selection kit Anti-Ly-6G MicroBeads UltraPure (Miltenyi Biotec, Cologne, Germany), following the manufacturer’s instructions.

The purity of isolated neutrophils verified by flow cytometry (FACSCanto II; BD Biosciences) using Ly6G-allophycocyanin Ab (BD Biosciences) showed a percentage of around 95.5 and 98.9% (data not shown). After the neutrophil purification, we counted the cells using trypan blue and verified that the cell viability was greater than 95% (data not shown). In this study, we used the following nomenclature to designate the Rictor-deleted neutrophils (NeuRic Δ/Δ) and the normal control (NeutRic fl/fl).

Genomic DNA from bone marrow–purified neutrophils was obtained using the Genomic DNA Extraction Mini Kit (MACHEREY-NAGEL, Germany), following the manufacturer’s instructions. The mice used in this study (Rictor^tm1.2Mgn) are described to have exon 3 removed from the locus, leaving a loxP site and an frt site, on expression of Cre recombinase in the germline. To detect the removal of exon 3 in the Rictor gene, we used the following primers: Forward 5′-ACTGAATATGTTCATGGTTGT-3′, Reverse 5′-GAAGTTATTCAGATGGCCCAG-3′ to detect rictor (mutant band–Loxp sequence inserted: 554 bp; wild type band: 466 bp); and internal positive control (324 bp) Forward (oIMR7338) 5′-CTAGGCCACAGAATTGAAAGATCT-3′, Reverse (oIMR7339) 5′-GTAGGTGGAAATTCTAGCATCATCC-3′. We used 10 ng of genomic DNA for each sample (Supplemental Fig. 1A, 1B).

For the in vitro and in vivo experiments, we used E. coli (ATCC 25992) or Pseudomonas aeruginosa (ATCC 27853). First, we made a curve of bacterial growth for each bacterium to use bacteria in the exponential phase of growth (100% viability). For the inoculum, we took 10 µl from an aliquot of the bacteria, previously frozen in glycerin, and kept it at −80°C. We added this 10 µl to 5 ml of Luria broth (LB) and incubated it for 16 h at 37°C in a shaker at 200 rpm. The next day, we performed the subculture adding 10 µl of the inoculum to fresh 5 ml of LB and incubated it for 4 h at 37°C in a shaker at 200 rpm. After this incubation, we measured the absorbance in a spectrophotometer at 600 nm. The absorbance value was used to calculate the number of bacteria using our bacterial curve. The number of E. coli used to induce sepsis in the mice was 1.5 × 10⁸ CFUs in 200 µl of sterile PBS. In our in vitro experiments, the multiplicity of infection (MOI) used was 50 for E. coli and 20 for P. aeruginosa.

To determine the neutrophils’ chemotactic capacity, we used a Transwell system (Corning), according to the manufacturer’s instructions. We used an insert in which the membrane had a 3 μm pore size and 10 nM fMLF as a chemoattractant. We added 1 × 10⁵ bone marrow–purified neutrophils in 100 μl of RPMI 1640 (Thermo Fisher Scientific) on the top of the inserted. At the bottom (in contact with the plate), we added 590 μl of RPMI 1640 and 10 μl of fMLF (10 nM) (Sigma-Aldrich) or vehicle (PBS-diluted DMSO). We incubated the cells in the Transwell system at 37°C and 5% CO₂ for 1 h. After incubation, we counted the cells that migrated to the bottom of the plate in five random microscope fields.

We plated 1.5 × 10⁵ neutrophils on an eight-well glass chamber slide (Nunc Lab-Tek; Thermo Fisher) in HBSS and allowed them to attach to the glass chamber slide for 30 min at 37°C. Subsequently, we changed the medium for PBS-glucose (10 mM Na₂HPO₄ PBS; 2 mM KH₂PO₄; 137 mM NaCl, 1 mM CaCl₂, 0.5 mM MgCl₂, 1 g/l glucose) and stimulated or not the neutrophils with 1 µM fMLF (Sigma-Aldrich). We incubated the chamber slide for 3 min at 37°C. Right after, we fixed samples with 4% paraformaldehyde for 15 min and washed them with PBS. Subsequently, we permeabilized the cells with 0.1% Triton-X (Sigma-Aldrich) for 15 min and blocked them with 2% BSA (Sigma-Aldrich) for 1 h. We removed the blocking solution and added a PBS+2% BSA solution containing phalloidin (1:40; Thermo Fisher), F-actin dye, and the DNA dye, Hoechst (1:500; Thermo Fisher), and incubated them for 45 min protected from light. We used Vectashield as a mounting medium, covered the slides with a cover glass, and stored them at 4°C for further analysis in a Zeiss LSM 780-NLO confocal microscope (Carl Zeiss, Jena, Germany).

To assess the microbicidal capacity of the neutrophils, we incubated 2 × 10⁵ neutrophils purified from bone marrow with E. coli (MOI = 50) in a 96-well plate at 37°C for 1 h in PBS-glucose. To assess the E. coli killing by phagocytosis, after incubation, we carefully removed and discarded the supernatant. To ensure that no bacteria were attached to the neutrophils or on the plate bottom, we incubated the neutrophils with a solution containing 120 µg/ml gentamicin in PBS for 30 min. After this incubation, the solution was removed, and the neutrophils were washed with PBS and lysed with a 0.1% Triton-X in PBS. The final volume of this solution (100 µl) was used to perform serial dilutions that were plated on LB agar and incubated at 37°C overnight. The CFUs were counted the day after. To evaluate the total bacterial killing, we used the same methodology, but instead of removing the supernatant and adding gentamicin, we lysed the cells with 0.1% Triton-X in PBS right after the 1-h incubation.

To quantify the phagocytic activity of neutrophils, we used zymosan particles (Sigma-Aldrich) resuspended in PBS. The particles were mixed well and washed three times with PBS-glucose. We stored them in 1 ml of PBS in a final stock concentration of 1 × 10⁷ particles/ml. For the experiment, we attached 1 × 10⁵ neutrophils on an eight-well glass chamber slide (Nunc Lab-Tek; Thermo Fisher) treated with poly-l-lysine for 1 h. After the attachment, we incubated the neutrophils with 5 × 10⁵ zymosan particles (1:5) for different time points (15, 30, 60 min) at 37°C. After the incubation, we removed the supernatant, and the cells on the slides were fixed and stained with H&E. For the internalized zymosan particles count, we used a light microscope and counted ∼100 neutrophils, and the particles internalized by them. The phagocytic index was calculated by the following ratio: number of phagocytosed particles/number of cells.

For the electron microscopy, 1 × 10⁷ neutrophils were incubated with E. coli (MOI = 50) for 2 h at 37°C in PBS-glucose. Subsequently, cells were washed twice with PBS and centrifuged at 300 × g for 10 min. The cell pellet was resuspended in a fixation solution containing 2.5% glutaraldehyde and centrifuged at 10,000 rpm for 10 min. This pellet was embedded in a specific resin and carefully placed on copper membranes for subsequent analysis by transmission electron microscopy, using a FEI Tecnai G20 200kv microscope (FEI Company).

To quantify specifically the HOCl generated by neutrophils after their activation by E. coli or fMLF, we used a specific fluorescent probe R19S (22). We incubated in PBS-glucose 1 × 10⁶ neutrophils with 10 µM R19S probe in a 96-well plate and stimulated or not the neutrophils with E. coli (MOI = 50) or fMLF (10 nM). The kinetics were performed for 4 h by the BioTek microplate reader (Synergy Hybrid) at 37°C. The fluorescence emitted by the oxidized probe by the HOCl was observed at a wavelength of 515/545 nm (22, 23). Baseline fluorescence was discounted using the control group (nonstimulated cells).

To quantify the superoxide generated by neutrophils after their activation by E. coli, we used a fluorescent dihydroethidium (DHE) probe. We incubated in PBS-glucose 1 × 10⁶ neutrophils with 10 µM DHE in a 96-well plate and stimulated or not the neutrophils with E. coli (MOI = 50). The kinetics were performed for 4 h by the TECAN microplate reader (Tecan, Männedorf, Switzerland). In this method, DHE oxidation generates two different compounds, ethidium (E+), which is the product of the reaction with two-electron oxidants, and 2-hydroxyethide (2-OH-E+), a product generated specifically from the probe’s reaction with the radical anion superoxide. The detection of 2-OH-E+ can be discriminated from E+ by exciting the samples at λ_exc = 396 nm with λ_em = 579 nm (23, 24). Basal fluorescence was discounted using the control group (nonstimulated neutrophils).

For the protein analysis, we stimulated 2 × 10⁶ bone marrow–purified neutrophils with 30 nM fMLF, 2 µg/ml LPS, or vehicle (PBS-diluted DMSO) and incubated them for 15 min or 2 h in PBS-glucose at 37°C. After incubation, neutrophils were lysed by the radioimmunoprecipitation assay (RIPA) buffer. Approximately 100 μg of total protein was diluted in a sample buffer (Laemmli; Bio-Rad Laboratories, Hercules, CA) containing 20 mg/ml DTT (Sigma-Aldrich). Proteins were denatured by heating (5 min at 95°C) and then separated by electrophoresis in 12% polyacrylamide gel. Subsequently, the proteins were transferred to a nitrocellulose membrane and blocked for 1 h in 5% milk dissolved in TBS containing 0.05% of Tween 20 for further incubation with primary Abs overnight. After this incubation, the membrane was washed with TBS containing 0.05% of Tween 20 and incubated for 1 h with secondary Abs conjugated with HRP. The molecular mass of protein was determined by the comparison with the migration pattern of Precision Plus Protein Prestained Standards Dual Color (Bio-Rad Laboratories).

The primary Abs used were anti-pAktSer473, anti-Akt (total) (1:1000; Cell Signaling Technology, MA), anti-β-actin (1: 10,000; Sigma-Aldrich), and anti-MPO (1:1000; Abcam).

To quantify the NO generated by neutrophils, we used a specific fluorescent probe, 4,5-diaminofluorescein diacetate (Thermo Fisher), according to the manufacturer’s instructions. In brief, 3 × 10⁵ neutrophils were incubated in PBS-glucose, with 5 µM 4,5-diaminofluorescein diacetate for 20 min at 37°C, and washed twice with PBS to remove the excess of the probe that did not enter the cells. The neutrophils were then resuspended in PBS-glucose and plated in a 96-well plate. The cells were stimulated or not with E. coli (MOI = 50) and placed at the Synergy microplate reader at 37°C. The fluorescence was recorded for 1 h with a wavelength of 485/535 nm. Basal fluorescence was discounted using the control group (nonstimulated neutrophils).

For the NETs quantification assay, we used Sytox Green (Life Technologies), a cell-impermeable fluorescent DNA dye. First, we incubated the neutrophils with Sytox Green (3 µM) and plated 2 × 10⁵ neutrophils per well in a 96-well plate in a PBS-glucose medium. Subsequently, we stimulated or not the neutrophils with P. aeruginosa (MOI = 20), PMA (100 nM, positive control), or vehicle (PBS-diluted DMSO). The plate was kept inside the Synergy microplate reader (Mx; BioTek) at 37°C, and the Sytox Green fluorescence intensities were measured at 30-min intervals for 240 min.

After cell isolation, we plated 2 × 10⁵ neutrophils per well on poly-l-lysine (100 μg/ml; Sigma-Aldrich)-coated Seahorse XFe96 Cell Culture Microplate (Agilent Technologies) in HBSS for 1 h at 37°C in CO₂-free incubator. Subsequently, the medium was changed into RPMI 1640 glucose and bicarbonate-free (Agilent Technologies). The glycolytic capacity of neutrophils was assessed by the extracellular acidification rate using the glycolytic stress test from XFe96 Extracellular Flux Analyzer (Agilent Technologies). The following reagents were injected at specific time points: 10 mM glucose (Sigma-Aldrich), 1 µM oligomycin (Sigma-Aldrich), and 100 mM 2-deoxyglucose (Agilent Technologies).

For the mitochondrial function assessment, we measured the oxygen consumption rate by the mitochondrial stress test using the XFe96 Extracellular Flux Analyzer (Agilent Technologies). We plated 2 × 10⁵ neutrophils per well on poly-l-lysine (100 μg/ml; Sigma-Aldrich)-coated Seahorse XFe96 Cell Culture Microplate (Agilent Technologies) in HBSS for 1 h at 37°C in CO₂-free incubator. Subsequently, the medium was changed into an XF assay medium RPMI 1640 supplemented with 10 mM glucose (Agilent Technologies). The following reagents were injected at specific time points: oligomycin (1 mM; Sigma-Aldrich), FCCP (5 mM; Sigma-Aldrich), rotenone (1 µM; Sigma-Aldrich), and antimycin A (1 µg/ml; Sigma-Aldrich).

We incubated 2 × 10⁶ neutrophils in 100 μl of PBS-glucose for 2 h at 37°C. The supernatant was collected after cell centrifugation and stored at −80°C. We measured the lactate in the supernatant using the Lactate Liquiform colorimetric kit (Labtest Diagnóstica, Lagoa Santa, Brazil), following the manufacturer’s instructions. In brief, we mixed 10 µl of the supernatant with the reagents provided in the kit and incubated them for 5 min in a water bath at 37°C. The sample absorbance was measured using a microplate reader (Synergy Mx; BioTek). The wavelength used was 550 nm, and the lactate concentration was calculated using the following formula: [Lactate] = (sample absorbance/standard absorbance) × 40.

Before performing the glucose uptake assay, we stained 4 × 10⁵ bone marrow–purified cells by Histopaque-1077 with Live/Dead-AmCyam dye (Thermo Fisher) and with the surface neutrophil marker anti-Ly6G-allophycocyanin (BD Biosciences) and anti-CD11b-phycoerythrin (BioLegend) for 30 min at 4°C. After washing the cells with PBS, we resuspended them in 100 µl of RPMI 1640 bicarbonate-free supplemented with 5 mM glucose and 5 µM fluorescent glucose analogue, 2(N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl) amino)-2-deoxyglucose, for 20 min at 37°C. After incubation, the cells were washed and resuspended in PBS containing 2% FBS and immediately acquired in a flow cytometer (FACSCanto II; BD Biosciences).

For the confocal live images, we first attached 2 × 10⁵ bone marrow–purified neutrophils in the µ-Dish 35 mm, using the culture-insert 2 wells (Ibidi, Gräfelfing, Germany) to record images from both neutrophils (NeutRic fl/fl and NeutRic Δ/Δ). To attach the cells, we treated the plate with poly-l-lysine (0.1 mg/ml) (Sigma-Aldrich, St. Louis, MO), following the manufacturer’s instructions. We left the neutrophils in HBSS resting for 1 h at 37°C to attach them to the plate. After, we preincubated them with 100 nM LysoTracker Deep Red (Invitrogen Life Technologies, Carlsbad, CA) for 15 min, followed by HBSS washing. Just before recording in the confocal microscope, we changed the medium to PBS-glucose and added the pHrodo Green E. coli BioParticles. This reagent is an E. coli conjugated with a molecule that fluoresces brightly green at acidic pH, such as in phagosomes. The lysotracker is a dye for labeling and tracking acidic organelles in live cells, such as the neutrophil granules. Therefore, these markers were used to observe the colocalization of the phagosome (pHrodo Green E. coli) and the acidic granules (LysoTracker Deep Red).

Images from the experimental groups were digitized with identical acquisition settings and analyzed. The interval of acquisition of the images was between 2.5 and 10 min (overnight). A threshold paradigm was used for the normalization and quantification of the immunofluorescence signal. Here we used Li’s Minimum Cross-Entropy thresholding method (25). The images treated were then colocalized using the “colocalization” plugin on the ImageJ software. This plugin initially generates an 8-bit image with only the colocalized points (image available by validating colocalized points 8-bit). Two points are considered as colocalized if their respective intensities are strictly higher than the threshold of their channels [which are 50 by default: Threshold channel 1 (0–255)], and if their ratio (of intensity) is strictly higher than the ratio setting value [which is 50% by defect: ratio (0–100%)]. The images were quantified in percentage (%) of area; the result was divided by the number of cells per field and then used for plotting the graphs (quantification of colocalization). Images were acquired on a Zeiss LSM 780 laser scanning confocal microscope and quantified by the ImageJ software (National Institutes of Health, Bethesda, MD).

First, we stimulated or not 1.5 × 10⁶ bone marrow–purified neutrophils with LPS (2 µg/ml) in PBS-glucose. After stimulation, neutrophils were prepared for metabolite extraction. They were washed twice with cold PBS, and 1 ml of extraction solution (acetonitrile-isopropanol-MilliQ H₂O 3:3:2 [v/v]) was added into the cells and transferred to a Precellys tube with ceramic beads, where the cells were homogenized with two cycles of 1500 rpm for 20 min. The homogenized solution samples were centrifuged at 15,800 × g for 5 min at 0°C. The supernatant (900 μl) was transferred to a new tube, spiked with 5 μl of D27 myristic acid (3 mg/ml), and dried in SpeedVac (Thermo Fisher) for 18 h. Next, for the derivatization phase of the protocol, dried metabolites were solubilized in 20 μl of methoxylamine diluted in pyridine (40 mg/ml), spiked with 3 μl of fatty acid methyl esters, and incubated at 25°C for 16 h at 650 rpm agitation. Then, 90 μl of MSTFA was added with 1% TMCS to each sample. This solution was incubated for 1 h at 25°C, 650 rpm agitation, followed by centrifugation 15,800 × g for 10 min at room temperature. The supernatants (70 μl) were transferred to a glass insert and run on gas chromatography-mass spectrometry. Each sample ran for 24 h. Five samples plus a quality control were prepared every day. Samples were injected randomly in gas chromatography–mass spectrometry in triplicate. Targeted metabolomics was performed according to Calderón-Santiago et al. (26) with some modifications. One microliter of each sample was injected into an Agilent 7890B GC system in a splitless mode. A DB5-MS + 10 m DuraGuard capillary column was used at a helium gas rate of 0.89 ml/min. The column temperature was held at 60°C for 1 min and then increased to 310°C at a rate of 10°C/min for 37 min. The detector (5977A; Agilent Technologies) operated in the electron impact ionization mode (70 eV), and mass spectra were recorded after a solvent delay of 6.5 min. Mass spectrometry operated in a SIM mode at a scan speed of 1.562 u/s and dwell time of 50 ms/m/z. A quantifier and a qualifier m/z were monitored for the pyruvic acid. Data were analyzed on the Agilent MassHunter Quantitative Analysis. Each sample was analyzed in two technical replicates. The standard curve was run after and before the samples.

Flow cytometry experiments were performed on a FACSCanto II (BD Biosciences), and data analysis was performed using the FlowJo 9.5.3 software (Tree Star, San Carlo, CA). The markers used were Live/Dead-AmCyam dye (Thermo Fisher), anti-Ly6G-allophycocyanin (BD Biosciences), and anti-CD11b-phycoerythrin (BioLegend).

The blood glucose levels were measured by the blood glucose meter Accu-Chek (Roche, Basel, Switzerland).

The evaluation of kidney dysfunction was performed by measuring the plasma urea using the Labtest colorimetric kit (Labtest Diagnóstica), according to the manufacturer’s specifications. In brief, after collection in sodium heparin, the blood was centrifuged at 10,000 rpm for 10 min at 4°C. Plasma was collected and stored at −80°C. For the urea measurement, we used 10 µl of plasma and incubated it with the reagents provided by the kit for 5 min in a water bath at 37°C, followed by another incubation in a dry bath at 37°C for 5 min. The sample absorbance was measured using a microplate reader (Synergy Mx; BioTek) at 620 nm, and the urea concentration was calculated using the following formula: [Urea] = (sample absorbance/standard absorbance) × 70.

Kidney and liver were harvested from animals and stored at −80°C. For the RNA extraction, we used TRIzol reagent (Thermo Fisher Scientific), according to the supplier instructions, using the tissue homogenizer Precellys (Bertin, Montigny-le-Bretonneux, France). The preparation of cDNA was performed using 2 μg RNA, according to the manufacturer’s instructions (Integrated DNA Technologies, Coralville, IA). Subsequently, we quantified gene expression by quantitative PCR (qPCR). The detection of the genes of interest (tnf-α, il-1β, kc, mpo, kim-1, and hprt) was performed using TaqMan primers (Applied Biosystems), according to the manufacturer’s instructions. The gene amplification was performed using the QuantStudio (Applied Biosystems), and the quantification method used was the relative expression with hprt as a housekeeping control gene. The control group (LysMRic fl/fl nonseptic mice) was used as a calibrator for our samples.

The kit Mouse Cytokine CBA (Cytometric Bead Array) Assay (BD Biosciences) was used to quantify TNF-α, IL-6, IFN-γ, MCP-1, and IL-10 in the mouse serum. The test was carried out according to the manufacturer’s instructions. The samples were acquired at FACSCanto II (BD Biosciences) and analyzed using the FCAP Array software (BD Biosciences).

Respiratory exchange ratio, the volume of oxygen consumed, and heat generation were analyzed using an indirect open circuit calorimeter (Oxymax System; Columbus Instruments, Columbus, OH). One animal was kept per cage for metabolic measurements. Animals were isolated and monitored by the electrodes present in each box with water and food ad libitum. We acclimatized them to this system for 24 h before starting the experiment. The system was previously calibrated with a specific gas mixture, and we evaluated the metabolic parameters for 24 h. Subsequently, we injected E. coli i.p. at night (at 8 pm). We assessed the metabolic parameters during the sepsis period for 12 h.

For the evaluation of the CFUs in the peritoneal cavity, blood, liver, kidney, and lung of the animals after sepsis, we harvested the organs and the liquids in a laminar flow (sterile microenvironment). For the peritoneal lavage, we injected 5 ml of PBS and removed 1 ml. Of this 1 ml, we used 100 µl for serial dilutions. We took the blood from the orbital plexus, ∼500 µl, and used 100 µl for serial dilutions. For the organs, we removed a small portion of each organ, weighed it, and macerated it with 1 ml of PBS. We also used 100 µl to perform serial dilutions. We plated the dilutions on LB agar and incubated them overnight at 37°C. We normalized the number of bacteria by the organ weight for each animal (CFUs/g).

After sepsis induction by E. coli (1.5 × 10⁸), the mice were monitored hourly.

We manually curated the Gene Expression Omnibus (GEO) repository to find transcriptome datasets related to sepsis in human patients. Differential expression analysis was performed based on author-normalized expression values (log₂ scale) using the GEO2R software (https://www.ncbi.nlm.nih.gov/geo/geo2r/). Probes that matched the same gene symbol were collapsed by taking the one with the lowest p value. We selected one dataset for analysis (GSE64457), which contains transcriptomic data from cell sorting–isolated neutrophils from peripheral blood of 15 patients with septic shock presenting clinical features of sepsis-induced immunosuppression and 8 healthy individuals. Only patients alive at day 3 or 4 after the onset of shock were considered. The onset of septic shock was defined as the beginning of vasopressor therapy. We first filtered the differentially expressed genes with a cutoff of p < 0.001 from the list obtained through GEO2R. Then, using the Metascape software (27), we generated a bar graph of the enriched pathways based on the differentially expressed gene list previously obtained. Pathway and process enrichment analysis was carried out using the Kyoto Encyclopedia of Genes and Genomes Pathway database as a source. All genes in the genome have been used as the enrichment background. Terms with a p value <0.05, a minimum count of 3, and an enrichment factor >1.5 (the enrichment factor is the ratio between the counts observed and the counts expected by chance) were collected and grouped into clusters based on their membership similarities. More specifically, p values were calculated based on the accumulative hypergeometric distribution, and q values were calculated using the Benjamini-Hochberg procedure to account for multiple tests. κ scores are used as the similarity metric when performing hierarchical clustering on the enriched terms, and subtrees with a similarity of >0.3 are considered a cluster. The most statistically significant term within a cluster was chosen to represent the cluster. The remaining significant terms were then hierarchically clustered into a tree based on κ statistical similarities among their gene memberships. Then, a 0.3 κ score was applied as the threshold to cast the tree into term clusters. In the end, using transcriptional data of the neutrophils isolated from patients with sepsis obtained from GEO2R, we performed correlation analyses between rictor and several genes related to the regulation of glycolysis, such as Hexokinase-1 (hk-1), Lactate dehydrogenase-A (ldh-a), Phosphoglycerate mutase-1 (pgam-1), and Enolase-1 (eno-1).

Statistical analysis was performed using the GraphPad Prism software (San Diego, CA). Differences among groups were compared using the Student t test or one-way or two-way ANOVA with Bonferroni posttest. We also used log-rank (Mantel–Cox) test to analyze the survival curve. The differences observed were considered significant when p < 0.05 (5%).

Previous studies using neutrophil-like cell lines demonstrated that mTORC2 activation is involved in cytoskeleton rearrangement (16, 17). To confirm this finding and investigate it in primary Rictor-deleted neutrophils, we first evaluated the mTORC2 activity in purified neutrophils. We showed that Akt phosphorylation at serine 473 (mTORC2 site of activity) was reduced in NeutRic Δ/Δ after fMLF or LPS stimulation (Fig. 1A, 1B). Then, we performed a chemotaxis assay using purified Rictor-deleted neutrophils and fMLF as a chemoattractant. Corroborating the literature, chemotaxis was strongly impaired in NeutRic Δ/Δ. (Fig. 1C). We also analyzed cytoskeleton rearrangements using confocal microscopy, and NeutRic Δ/Δ showed reduced cell membrane protrusions (lamellipodia) emission compared with NeutRic fl/fl after fMLF stimulation (Fig. 1D). These results suggest that the absence of Rictor in neutrophils affects the lamellipodia formation, impairing their chemotaxis induced by fMLF. We then evaluated whether the absence of Rictor would also impact their microbicidal and phagocytic capacity.

Apart from the cytoskeleton rearrangement, the importance of mTORC2 in other neutrophil effector functions has not been fully understood yet. Because the most important function in neutrophils is their capacity to kill pathogens, we aimed to investigate the importance of mTORC2 for neutrophil’s killing capacity. As shown in (Fig. 2, NeutRic Δ/Δ incubated with E. coli exhibited an increased CFU in both experimental approaches, intracellular killing (Fig. 2A) and intracellular and extracellular killing (Fig. 2B), compared with NeutRic fl/fl. To investigate why the microbicidal capacity was impaired, we first evaluated the phagocytosis proficiency. To do so, we incubated purified neutrophils with zymosan and counted internalized particles (Fig. 2C, 2D). We also performed electron microscopy to verify whether the neutrophils indeed internalized E. coli (Fig. 2E). Both quantitative and qualitative methods showed that the absence of Rictor did not affect the phagocytic capacity of neutrophils (Fig. 2C). However, the electron microscopy images showed that the E. coli electron density was increased in the NeutRic Δ/Δ, suggesting that the E. coli was not being effectively digested by neutrophils in the absence of Rictor (28). Another aspect of neutrophil function is their capacity to produce oxidants during the oxidative burst, so we questioned whether this would be impacted by Rictor as well.

Understanding that the microbicidal capacity of neutrophils relies on oxidant production, we first investigated HOCl production because of its powerful microbicidal capacity. We observed a decrease in HOCl production in Rictor-deleted neutrophils (Fig. 3A, 3B). Then, we investigated the first oxidant produced by NADPH oxidase during the oxidative burst, the radical anion superoxide. Surprisingly, Rictor deficiency in neutrophils did not affect the production of this specific oxidant (Fig. 3C, 3D), showing that the cascade of events was not completely altered. Because MPO is the limiting enzyme to HOCl production, we explored its expression in both neutrophils, NeutRic Δ/Δ and NeutRic fl/fl, and did not observe any significant differences (Fig. 3E, 3F). These results indicate that the microbicidal capacity in NeutRic Δ/Δ was impaired because of a decrease in HOCl production. To analyze another aspect of microbicidal capacity, we evaluated NETosis.

Since the discovery of NETs, several studies have demonstrated the importance of oxidants, including HOCl, for their formation (29, 30). Moreover, several signaling pathways are activated to promote their release. To verify the importance of mTORC2 to NETs formation, we performed a quantitative NETs assay using PMA (Fig. 4A, 4B) or P. aeruginosa (Fig. 4C, 4D). We observed a reduction of NETs in NeutRic Δ/Δ compared with NeutRic fl/fl in the quantitative analysis (Fig. 4A–D) and in the images (taken 4 h after the stimulation of neutrophils) (Fig. 4E). Another aspect of neutrophil biology that has been recently associated with its effector function is the cellular metabolism. Because mTOR itself is a master regulator of the bioenergetic status of mammalian cells, we further evaluated the metabolic landscape in these cells.

A growing body of evidence has shown that neutrophils are not purely glycolytic, and their metabolism can be shaped not only by host metabolites but also by those derived from infectious agents (10, 12). To investigate the importance of mTORC2 on neutrophil metabolism, we observed key metabolites and oxygen consumption, relating them to aerobic glycolysis or oxidative phosphorylation. In basal conditions (nonstimulated cells), we observed that the absence of Rictor did not affect the metabolism of neutrophils (Fig. 5A, 5B). However, after stimulation, NeutRic Δ/Δ showed a higher lactate production (Fig. 5C), suggesting higher levels of aerobic glycolysis, even though showing a lower glucose uptake compared with NeutRic fl/fl (Fig. 5D). An important recognized hub on glucose metabolism is the fate of glycolysis-generated pyruvate, and to evaluate the intracellular concentration of this metabolite, we performed targeted metabolomics. We observed that pyruvate concentration was decreased in NeutRic Δ/Δ (Fig. 5E), also corroborating the idea that these neutrophils shift toward lactate production using aerobic glycolysis (Fig. 5C). In this sense, these results suggest that the absence of Rictor in neutrophils affects their mitochondrial metabolism, exacerbating the glycolytic pathway. The metabolic analysis is important because the metabolism is intertwined with the neutrophil effector functions, such as phagosome activation.

The intracellular killing of pathogens in phagocytes involves the fusion of lysosomes containing antimicrobial substances with phagosomes. To find a mechanism by which the HOCl (Fig. 3A, 3B) and NO (Supplemental Fig. 1C, 1D) production were decreased in NeutRic Δ/Δ, we investigated the importance of Rictor for the granules trafficking, specifically the fusion with the phagosome. Using confocal microscopy, we assessed the fusion between the specialized lysosome and the phagosome (colocalization in orange-yellow), qualitatively by images (Fig. 6A) and video (Supplemental Video 1), and quantitatively by measuring the colocalization (Fig. 6B, 6C). We observed a decrease in colocalization (fusion of granules) in Rictor-deleted neutrophils. This finding suggests that HOCl and NO decreased production is due to an improper fusion of granules with the phagosome (Fig. 6B, 6C), which impairs the microbicidal capacity in NeutRic Δ/Δ. Having extensively characterized the impact of Rictor for neutrophil function in isolated cells, we finally wanted to understand whether the several impairments could impact an in vivo sepsis model.

To identify the sepsis outcome in myeloid-Rictor deficient mice (LysMRic Δ/Δ), we first evaluated the animal metabolic state by indirect calorimetry (Fig. 7A–C) and verified that the oxygen consumption (Fig. 7B) and the heat generated (Fig. 7C) in septic LysMRic Δ/Δ mice were reduced compared with LysMRic fl/fl mice. We also evaluated metabolic parameters and observed that LysMRic Δ/Δ mice were more hypoglycemic (Fig. 7D), and that urea (Fig. 7E) and TNF-α (Fig. 7F) plasma levels were higher compared with LysMRic fl/fl mice. Corroborating a worsening sepsis condition, the proinflammatory cytokines gene expression in liver and kidney, such as tnf-α (Fig. 7G, 7I) and il-1β (Fig. 7H, 7J), were increased in LysMRic Δ/Δ. We also evaluated a marker of kidney injury, the kim-1 gene expression, and it was increased in LysMRic Δ/Δ mice (Fig. 7K). Although these parameters indicated a poor outcome in LysMRic Δ/Δ, the bacterial load in target sepsis organs is crucial to evaluate the disease.

The recruitment of neutrophils from the circulation into the inflammatory site during a bacterial infection is a crucial step for the host to restrain and eliminate the pathogen. After 12 h of E. coli injection in mice, we evaluated the bacterial load in target sites, such as the peritoneal cavity, blood, and liver (Fig. 8A–C). Corroborating our in vitro results, the CFU in these sites was increased in LysMRic Δ/Δ mice. To investigate whether it was due to an improper neutrophil migration to the site of infection, we inoculated into the peritoneal cavity a lower number of E. coli and verified neutrophil migratory capacity after 2 h. Unlike the stimulus with fMLF in vitro, in which the absence of Rictor impaired migration of neutrophils (Fig. 1C, 1D), E. coli could induce equally the recruitment of neutrophils to the peritoneal cavity in both mice, LysMRic Δ/Δ and LysMRic fl/fl (Fig. 8D). To analyze the status of activation of these neutrophils, we divided these cells into two populations, CD11b^high (more activated) and CD11b^low (less activated). We showed that neutrophil activation in LysMRic Δ/Δ mice was impaired compared with LysMRic fl/fl mice (Supplemental Fig. 1E, 1F). Furthermore, the bacterial load 2 h after the E. coli injection was also increased in LysMRic Δ/Δ mice (Supplemental Fig. 1G), confirming that the microbicidal capacity is impaired in vivo and is not dependent on the number of neutrophils (Supplemental Fig. 1F). Because homozygous LysMCre mice may have affected lysozyme M production, we performed a microbicidal capacity assay using the LysMCre^+/+ and Rictor floxed mice (LysMRic fl/fl) as controls to confirm the absence of this enzyme was not playing a major role in our model. The bacterial burden was increased in the peritoneal cavity of LysMRic Δ/Δ mice, but not in the LysMRic fl/fl or LysMCre^+/+ mice (Supplemental Fig. 1H). We also showed that the neutrophil migration to the peritoneal cavity is similar among the groups (Supplemental Fig. 1I).

Despite that our results suggest a poor sepsis outcome in the LysMRic Δ/Δ mice, no statistical difference was seen in the survival rate between LysMRic Δ/Δ and LysMRic fl/fl mice (Fig. 8E) in this sepsis model.

Because we observed an increase in aerobic glycolysis in Rictor-deleted neutrophils (Fig. 5C), we evaluated its expression in neutrophils isolated from patients with sepsis, and we performed a correlation analysis between this gene and key regulators of the glycolytic pathway using a publicly available dataset (GEO: GSE64457). At first, we observed an enrichment of the autophagy (including mTOR pathway) and HIF-1 signaling pathways when comparing differentially expressed genes between neutrophils isolated from patients and healthy individuals (Fig. 9A), which suggests a possible modulation of neutrophil functions by the mTOR pathway during human sepsis and induction of glycolytic metabolism in these cells, because HIF-1α is a key inductor of glycolysis in activated neutrophils (31, 32). Then, we assessed the expression of rictor in neutrophils isolated from patients with sepsis, which is downregulated when compared with cells isolated from healthy control subjects (Fig. 9B). We also observed significant negative correlations between rictor and several genes related to the glycolysis pathway (Fig. 9C–F), suggesting that the downregulation of Rictor in isolated neutrophils from human patients would lead to increased glycolytic metabolism, which corroborates our data from the in vitro experiments.

On infection, neutrophils control their metabolic processes to adapt and optimize the effector functions according to their needs and the nutrients available (9, 10). In this sense, cell function and metabolism are intertwined. In this study, we evaluated the importance of mTORC2 activation for neutrophil effector functions by using a genetic deletion of the mTORC2-scaffold protein Rictor in these cells. We demonstrated that Rictor is crucial for phagosome-granules fusion and glycolytic metabolism, leading to a proper HOCl and NETs production and, consequently, an effective microbicidal capacity.

The mTOR pathway is one of the most studied metabolic signaling pathways in immune cells due to its activation by TLRs. Several studies demonstrated that LPS activates mTORC2 in monocytes, macrophages, and dendritic cells (13–15). Furthermore, in macrophages, mTORC2 plays a role in controlling excessive inflammatory response once its absence leads to an exacerbated inflammatory immune response (33). However, little is known about the importance of mTORC2 in neutrophils. Studies demonstrated that G-coupled protein receptor ligands activate mTORC2 in neutrophil-like cells, such as HL-60 or PLB-985, promoting F-actin– and Myosin II–dependent cytoskeleton rearrangement (16, 17) and also limiting actin assembly (negative feedback), assuring proper chemotaxis (18). These studies used shRNA to delete Rictor in neutrophil-like cells and showed a satisfactory degree of deletion. In our work, we checked the degree of Rictor deletion by the PCR of the rictor exon 3 and the phosphorylation of Akt at serine 473 (mTORC2 site of activity). We verified that the activity of mTORC2 was abrogated in Rictor-deleted neutrophils (NeutRicΔ/Δ). So, corroborating the literature, we confirmed the impairment of migratory capacity in Rictor-deleted neutrophils through an fMLF chemoattractant gradient (16, 17).

The cytoskeleton is an essential component in neutrophils, not only for chemotaxis but also for phagocytosis and vesicles/mitochondria trafficking (34). Although the cytoskeleton rearrangement is crucial for phagocytosis, the signaling pathways involved in the pathogen engulfment are different from those that activate the cytoskeleton dynamics for cell migration. Even in the same cell function, different receptor ligands activate distinct signaling pathways. As previously mentioned, studies showed that chemotaxis by fMLF or leukotriene B₄ is impaired in Rictor-deleted neutrophils (16, 17). However, the same studies also showed that chemotaxis triggered by granulocyte macrophage-colony-stimulating factor occurs normally in the absence of Rictor but is dependent on mTORC1 (16, 17, 35). In this sense, we demonstrated that the absence of Rictor did not affect the phagocytic capacity of neutrophils, despite research showing that Rictor is involved in cytoskeleton dynamics. The E. coli phagocytosis, for example, involves several receptors, such as Fc receptors and complement receptors, with the assistance of TLR4, and signaling pathways capable of promoting the engulfment of this bacterium (34). Although phagocytosis is not affected in the absence of Rictor, the microbicidal capacity was significantly reduced in NeutRic Δ/Δ compared with NeutRic fl/fl. The electron microscope images revealed that E. coli electron density was higher in NeutRic Δ/Δ than in NeutRic fl/fl, indicating an impairment in E. coli degradation (28), corroborating our killing assay.

The microbicidal capacity of neutrophils relies on the production of oxidants. Among the different types of oxidants produced by neutrophils, HOCl is the most potent oxidant against internalized pathogens (36, 37). In this sense, we suggested that the killing capacity of Rictor-deleted neutrophils was affected because of reduced HOCl production by these neutrophils. Moreover, the release of oxidants by neutrophils is associated not only with the killing of pathogens but also with NET formation (6, 38, 39). Some studies showed the importance of specific oxidants for NET release, such as H₂O₂ and HOCl (29, 30). It was found that HOCl alone can restore NET formation in patients with chronic granulomatous disease (29). These patients do not have functional NADPH oxidase and do not produce NETs, being susceptible to recurrent infections (29), thus demonstrating the importance of NETs in controlling infectious agents. Although the importance of NETs during infection is well known, the mechanisms and signaling pathways involved in its release remain unclear. Human neutrophils treated with mTOR inhibitors showed a decrease in NET formation, but not in their phagocytic killing capacity. The authors suggested that it was due to a mechanism involving the regulation of HIF-1α by the mTOR pathway (40). In this study, we suggest that mTORC2 activation is important for NET formation. However, the approach used to quantify NETosis, the Sytox Green nucleic acid stain, did not differentiate the type of cell death. Although the shape of the curve obtained after PMA stimulation corresponds to NET formation (38), we cannot assure that mTORC2 is involved only in the generation of NETs. It could participate in other types of cell death, such as late apoptosis or necrosis, for instance.

Because the absence of Rictor affects some neutrophil effector mechanisms and the metabolism is associated with these functions, we analyzed the metabolic profile of neutrophils.

Neutrophils usually generate energy through aerobic glycolysis, and mitochondria do not contribute effectively to ATP generation in this cell (41). However, mitochondrial activity is not only associated with energy production. Studies demonstrated that neutrophils’ chemotaxis and oxidant production by NADPH oxidase are impaired when ATP synthase (mitochondrion complex V) is inhibited by oligomycin (42), showing the importance of mitochondria for neutrophil functions. After stimulation with fMLF or LPS, Rictor-deleted neutrophils were more glycolytic (increase in lactate production), despite their slightly decreased glucose uptake. After entering the neutrophils, the glucose can be directed to different destinations. The carbon flux can be directed via glucose-6-phosphate dehydrogenase to the pentose phosphate pathway, generating NADPH and the biosynthesis of nucleic acids (43, 44), or it can follow the glycolysis enzymatic reactions, generating pyruvate. This metabolite is a hub on glucose metabolism, and it can be redirected to generate lactate, or it can enter the mitochondria, fueling the TCA cycle (45, 46). Because we observed that the pyruvate concentration is lower in NeutRic Δ/Δ than in NeutRic fl/fl mice after LPS stimulation, we believe that Rictor-deleted neutrophils are prioritizing aerobic glycolysis (metabolizing pyruvate to produce lactate), whereas the NeutRic fl/fl are diverting the glucose to other pathways and directing part of the pyruvate to mitochondria.

These metabolic changes in Rictor-deleted neutrophils likely contribute to the dysfunction in their effector mechanisms once mitochondria play a role in oxidant production by NADPH oxidase and can also produce oxidants, such as H₂O₂ (43). However, the HOCl production is the most affected oxidant by the deletion of Rictor. Because we observed that MPO protein expression is not affected in Rictor-deleted neutrophils, we decided to investigate the phagolysosome formation. After pathogen engulfment and phagosome formation and maturation, lysosomes fuse with the phagosome, generating the phagolysosome (3), allowing the phagosome to acquire degradative properties by generating oxidants from oxygen and nitrogen species by oxidative burst (3). The enzymes that generate HOCl and NO, MPO, and inducible NO synthase, respectively, are present in the azurophilic granules, also called specialized lysosomes in neutrophils (47). The mechanisms that coordinate the phagolysosome generation, such as the membrane traffic and release of granule contents in neutrophils are not fully understood due to the difficulties in manipulating neutrophils. In this sense, we demonstrated by colocalizing the phagosome and neutrophil’s acidic vesicles that Rictor is crucial for a proper fusion of granules in pathogen-containing phagosomes, leading to an impairment in HOCl and NO production in Rictor-deleted neutrophils.

Aiming to verify the impact of Rictor deficiency in neutrophils in vivo, we chose an acute monomicrobial E. coli–sepsis model to link our in vitro and in vivo findings. Furthermore, E. coli is one of the most commonly isolated bacteria in the blood culture from patients with sepsis (48). In this sense, we decided to induce severe sepsis aiming to visualize the neutrophil performance in this model. We evaluated different parameters by indirect calorimetry, which demonstrated that our sepsis model was severe and induced septic shock. The calorimetry is based on the exchange of respiratory gases to assess the animal’s metabolic state (49, 50). As expected, during sepsis, the RER value was around 0.8, indicating mice were metabolizing lipids, instead of carbohydrates, as an energy source, which causes weight loss (51). Although no difference between LysMRic Δ/Δ and LysMRic fl/fl was observed in RER, the oxygen consumption and heat were lower in LysMRic Δ/Δ mice. During sepsis, the volume of oxygen consumed decreases because animals reduce their activities and stay prostrated (50). The heat is a measurement of the energy used to maintain body temperature, which is regulated by the animal’s movement and also by the brown adipose tissue (52). These two parameters in calorimetry suggested a poor sepsis prognosis in LysMRic Δ/Δ mice. Besides calorimetry, other parameters indicate the severity of sepsis, such as hypoglycemia, plasma urea, and proinflammatory cytokines. During sepsis, hypoglycemia is associated with septic shock (53), and high levels of urea in plasma are associated with acute kidney injury, one of the sepsis markers (54). Similarly, high levels of certain proinflammatory cytokines are markers of severity in sepsis, such as TNF-α and IL-1β (55). We verified that LysMRic Δ/Δ mice showed higher levels of systemic TNF-α and other proinflammatory gene expression cytokines in the liver and kidney than LysMRic fl/fl, demonstrating that LysMRic Δ/Δ mice were more inflamed than LysMRic fl/fl. Our findings in calorimetry and proinflammatory status corroborate another study that used the same mice lineage but induced endotoxemia by LPS (33). The authors focused on macrophages and discussed the importance of mTORC2 to control exacerbated inflammatory responses in these cells.

Corroborating our in vitro findings, LysMRic Δ/Δ mice showed a defect in the microbicidal capacity with an increased bacterial load in the peritoneal cavity, blood, and liver. Our results on microbicidal capacity among the LysMCre mice showed that the presence of the Cre enzyme in the lysozyme M promoter is not playing a major role in our model. Although the lysozyme M is important to kill different types of bacteria (56, 57), pathogens have evolved mechanisms to evade the killing by lysozyme. Some Gram-negative bacteria, such as E. coli and P. aeruginosa, are known to be resistant to this enzyme due to the expression of periplasmic Ivy proteins (lysozyme inhibitors) (58–60). E. coli also have enzymes that recycle peptidoglycan, contributing to envelope integrity and lysozyme resistance (58). Furthermore, contrary to the defect in neutrophils chemotaxis by fMLF, we observed that Rictor-deleted neutrophils could migrate to the peritoneal cavity through E. coli stimulation at the same level as the control group. Although the number of neutrophils was similar between mice, the CD11b marker was decreased in LysMRic Δ/Δ. The increase of CD11b expression suggests a higher neutrophil activation (61). In this sense, Rictor-deleted neutrophils were less activated but still contributed to a hyperinflammatory status, likely to compensate for the killing dysfunction. Although many parameters pointed out a poor sepsis outcome in LysMRic Δ/Δ mice, we did not observe any statistical difference in the survival rate in our model. However, several studies reported a failure in the migration and antimicrobial function of neutrophils in patients with sepsis and relate it to a poor sepsis outcome (62–64). To provide a translational perspective in our study, we used a public dataset and showed that the mTOR pathway may play a role in the regulation of neutrophils during sepsis in which Rictor gene expression is downregulated in neutrophils isolated from patients with sepsis, suggesting that mTORC2 could be less activated in these cells, thus aggravating sepsis. It has been shown in PBMCs that the shift from oxidative phosphorylation to glycolysis is important during sepsis, being related to a better outcome (65). However, it might not be true for neutrophil, which is already a glycolytic cell. In this study, we demonstrated that Rictor-deleted neutrophils have an exacerbated glycolytic metabolism. Corroborating this finding, Rictor expression in neutrophils from patients with sepsis was negatively correlated with genes involved in the glycolysis pathway.

Taken together, our results demonstrated that Rictor is crucial for granules fusion in neutrophils, allowing the proper oxidant production and microbial killing. Furthermore, the lack of Rictor affects neutrophil metabolism, leading to an exacerbated aerobic glycolysis. In our E. coli–sepsis model, supporting the results in isolated neutrophils, we also observed the impairment in activity and microbicidal capacity of neutrophils. Finally, bioinformatics analysis supported our findings and evidenced that rictor is downregulated in neutrophils isolated from patients with sepsis, being negatively correlated with genes controlling the glycolysis pathway. Collectively, our findings suggest a role for mTORC2 in neutrophil function and metabolism during sepsis.

We thank Dr. Alexandre Steiner for kindly providing the E. coli and P. aeruginosa strains. We are thankful to CEFAP-USP, especially Iuri Cordeiro Valadão, for the confocal microscopy assistance and video confections, and the Department of Cell Biology and Development at ICB-USP for the transmission electron microscopy assistance.

The authors have no financial conflicts of interest.