Peritoneal tissue-resident macrophages are metabolically poised to engage microbes using tissue-niche fuels - PubMed (original) (raw)
Peritoneal tissue-resident macrophages are metabolically poised to engage microbes using tissue-niche fuels
Luke C Davies et al. Nat Commun. 2017.
Abstract
The importance of metabolism in macrophage function has been reported, but the in vivo relevance of the in vitro observations is still unclear. Here we show that macrophage metabolites are defined in a specific tissue context, and these metabolites are crucially linked to tissue-resident macrophage functions. We find the peritoneum to be rich in glutamate, a glutaminolysis-fuel that is exploited by peritoneal-resident macrophages to maintain respiratory burst during phagocytosis via enhancing mitochondrial complex-II metabolism. This niche-supported, inducible mitochondrial function is dependent on protein kinase C activity, and is required to fine-tune the cytokine responses that control inflammation. In addition, we find that peritoneal-resident macrophage mitochondria are recruited to phagosomes and produce mitochondrially derived reactive oxygen species, which are necessary for microbial killing. We propose that tissue-resident macrophages are metabolically poised in situ to protect and exploit their tissue-niche by utilising locally available fuels to implement specific metabolic programmes upon microbial sensing.
Conflict of interest statement
The authors declare that they have no competing financial interests.
Figures
Fig. 1
Peritoneal tissue-resident macrophages are highly oxidative and are fuelled by glutamine. a Heat-map showing the log10 difference from the average metabolite peak areas from gas chromatography-mass spectrometry, which are associated with the citric acid cycle from peritoneal tissue-resident macrophages (pTRMØ, Res) and bone marrow-derived macrophages (BMDM). Data show 3–6 independent samples per group analysed by two-way ANOVA (interaction < 0.0001) with Sidak’s post-tests. b Schematic showing a simplified metabolic pathway detailing readouts for extracellular acidification rate (ECAR), oxygen consumption rate (OCR) and drugs, which inhibit these pathways, the red oval indicates the mitochondria. c Graph showing representative Seahorse mitochondrial stress tests, O = oligomycin (1.26 μM), F = FCCP (0.67 μM), R/AA = Rotenone (0.2 μM) + antimycin A (1 μM). d Bar graphs showing quantified protein-normalised ECAR and mitochondrial OCR from stress tests in c, combination of the two OCR parameters represents the maximum mitochondrial capacity. Data (n = 3 separate wells per group) represent three experiments, OCR data were analysed by paired two-way ANOVA (interaction p < 0.0001) with Sidak’s post-tests for the difference between pTRMØ and BMDM (reserve/ maximal mitochondrial capacity) or by Student’s _t_-test (ECAR). e Bar graph depicting the ratio of mitochondrial 16s deoxyribonucleic acid (DNA) vs β-actin DNA in three independent samples, data was not significant by Student’s _t_-test. f Bar graph showing the drop in maximum decoupled mitochondrial OCR after addition of drugs or withdrawal of fuels. Glucose dependence was assessed using 2-deoxyglucose (2-DG, 100 mM), pyruvate using UK-5099 (20µM), a mitochondrial pyruvate transport inhibitor, long-chain fatty acids by etomoxir (100 μM) and glutamine by 1 h starvation before reintroduction of glutamine (2 mM) or control media. g Stacked bar graph showing a breakdown of the effects of different metabolic fuels on maximum mitochondrial capacity from f. Data f and g (n = at least 3 separate wells per group) represent at least two experiments, and were analysed by two-way ANOVA ((G) interaction p = 0.02; (F) interaction p = 0.056, cell p = 0.017, fuel p < 0.0001) with Sidak’s post-tests. All error bars denote mean ± SEM
Fig. 2
Peritoneal tissue-resident macrophage mitochondria can be fuelled through abundant peritoneal glutamate. a Bar graph showing the effect of 5× diluted peritoneal fluid (1 mM glucose media lavage followed by cell and protein depletion) in comparison to 1 mM glucose media on mitochondrial parameters in peritoneal tissue-resident macrophages. Data were from one experiment (n = 5–6 separate wells per group) and analysed by paired two-way ANOVA (interaction p < 0.0001) with Sidak’s post-tests. b Heat-map showing the log10 difference from the average metabolite peak areas from gas chromatography-mass spectrometry analysis of the most enriched amino acids in peritoneal lavage fluid and the fatty acids in blood serum, glucose and glutamine act as a reference. Data show five independent samples that were analysed by two-way ANOVA (interaction p < 0.0001) with Sidak’s post-tests. c Stacked graph detailing the addition of glutamine (Q, 0.25 mM) and glutamate (0.5 mM) to peritoneal tissue-resident macrophages and resulting changes in mitochondrial parameters. Data are from the same experiment as a, but increases in mitochondrial function after glutamate and Q represent three independent experiments, data were analysed by two-way ANOVA (interaction p < 0.0001), Sidak’s post-tests are indicated within the bars for parameters vs 1 mM glucose and above representing the synergistic effect of glutamate and Q. All error bars denote mean ± SEM
Fig. 3
Peritoneal tissue-resident macrophages enhance their respiratory burst with glutaminolysis. a Line graph showing oxygen consumption rate (OCR) in peritoneal tissue-resident macrophages (pTRMØ). Arrow = zymosan (50 μg/ml), LPS (100 ng/ml) or control. b OCR changes in pTRMØ with washed particle- or soluble-zymosan components. Data were analysed by Student’s _t_-test. c Quantification of pTRMØ uptake of pHrodo zymosan (50 μg/ml,1 h). G = glucose (25 mM), Q = glutamine (2 mM). Data were not significant by one-way ANOVA (p = 0.0756) with Tukey’s post-tests. d OCR or extracellular acidification rate (ECAR) vs time in pTRMØ. Cells were starved from glutamine (Q) and cultured ± glucose before addition of Q or control (grey arrow). Zymosan = black arrow. Peak burst is quantified in the chart and analysed by one-way ANOVA (p < 0.0001), significant Tukey’s post-tests for Q addition are shown. e Graph showing the total oxygen consumption of pTRMØ after zymosan, quantified from OCR vs time (as in d), in the presence of 2-deoxyglucose (2-DG, 100 mM), dehydroepiandrosterone (DHEA,100 μM) or vehicle. Data were analysed by Student’s _t_-test. f Graph showing the percentage oxygen consumption of pTRMØ after zymosan in the presence of etomoxir (50 μM) or vehicle. Data are pool of two experiments (n = 6–7) and were analysed by two-way ANOVA (interaction p = 0.376, etomoxir p = 0.2588). g Relative total oxygen consumption after zymosan of pTRMØ pre-treated with glutaminolysis inhibitors bis-phenylacetamido-thiadiazolyl-ethyl sulphide (BPTES, 15 μM) and diazo-oxo-norleucine (DON, 10 mM), in the presence or absence of Q. Data were analysed by two-way ANOVA (interaction p = 0.0536) with Sidak’s post-tests. h Total oxygen consumption after zymosan of pTRMØ in the presence or absence of Q (4 mM) or glutamate (4 mM). Data were analysed by one-way ANOVA (p < 0.0001) with Tukey’s post-tests. i Human monocytes were pre-cultured for 48 h with recombinant human M-CSF (50 ng/ml) and interferon-γ (10 ng/ml). Oxygen consumption was quantified as in d and shows six separate wells per group pooled from two experiments, analysed by two-way ANOVA (interaction p = 0.1231), the *** depicts the significant effect of Q (p = 0.0009). Data (a–e, g, h) represent two experiments with at least three separate wells per group. All error bars denote mean ± SEM
Fig. 4
Glutamine does not support neutrophil respiratory burst. a Line plot of OCR vs time in peritoneal tissue-resident macrophages and bone marrow neutrophils. Arrow indicates the addition of zymosan (50 μg/ml). Data are from two separate experiments and represent at least three independent observations. b Bar chart showing the quantification of pHrodo zymosan (50 μg/ml) uptake. Data show three separate samples per group, represents two independent experiments and was not significant by Student’s _t_-test. c Line plot showing OCR vs time in neutrophils in the absence of glutamine (Q) and presence or absence of glucose (25 mM) for 90 min before addition of Q (2 mM) or control media (grey arrow). Zymosan = black arrow. d Peak burst quantification of plots from c and ECAR plots from the data in c. Data c, d represent at two independent experiments with three separate wells per group. Peak burst data were analysed by one-way ANOVA (p = 0.0262), Tukey’s post-tests for the addition of glutamine were not significant. e Total oxygen consumptions were quantified by area under the curve of OCR vs time as seen in c, in presence of 2-deoxyglucose (2-DG,100 mM) or vehicle control. Data (n = at least 3 separate wells per group) represent two independent experiments and were analysed by Student’s _t_-test. All error bars denote mean ± SEM
Fig. 5
Peritoneal tissue-resident macrophages enhance mitochondrial function during respiratory burst through a glutaminolysis enhancement of complex-II. a Peritoneal tissue-resident macrophage (pTRMØ) and neutrophil basal-OCR. Data show at least 10 wells per group and represents three observations. b Changes in mitochondrial parameters after zymosan (50 μg/ml) to pTRMØ. Data (n = 6 separate wells per group) represent three experiments analysed by paired two-way ANOVA (interaction p < 0.0001) with Sidak’s post-tests. c The maximal mitochondrial capacity from b. d Graph showing the total oxygen consumption of neutrophils after zymosan, ±VAS-2870 (10 μM). Data (9 wells per group from three experiments) were analysed by two-way ANOVA (interaction p = 0.8567), *** the significant effect of VAS-2870 (p = 0.0005). e Oxygen consumption of pTRMØ as in d. Data (6 wells per group from two experiments) were analysed by two-way ANOVA (interaction p = 0.138), * the significant effect of VAS-2870 (p = 0.0331). f A simplified diagram of the electron transport chain (ETC). Inhibitors are shown in red. g The effect of rotenone (200 nM) on total oxygen consumption of neutrophils after zymosan. Data show three wells per group and represents three experiments. h The total oxygen consumption of pTRMØ after addition zymosan in the presence of the ETC inhibitors: oligomycin (1.26 μM), rotenone (200 nM), thenoyltrifluoroacetone (TTFA, 2 mM), antimycin A (1 μM). Data show three wells per group, represents two experiments and were analysed by one-way ANOVA (p < 0.001) with Tukey’s post-tests vs control. i The effect of metabolic drugs f and 2-deoxyglucose(2-DG,100 mM) on the uptake of pHrodo zymosan(50 μg/ml,1 h). Data were analysed by one-way ANOVA (p < 0.001) with Tukey’s post-tests vs control. j Graph showing change in OCR after addition of indicated drugs f and atpenin A5 (1 μM). Post treatment (Post) is 90 min after zymosan, pre-treatment occurs before zymosan (Pre). Data show three wells per group, represents at least two experiments and were analysed by two-way ANOVA (interaction p < 0.0001) with Sidak’s post-tests. k The effect of a 5 h atpenin treatment on OCR in pTRMØ. l Graph showing total oxygen consumption of zymosan-induced pTRMØ after atpenin A5 ±glutamine. Data (6 separate wells per group) is a pool of two experiments and was analysed by two-way ANOVA (Interaction p = 0.0079) with Sidak’s post-tests. All error bars denote mean ± SEM, except c which shows the median, data in a, c, g, k were analysed by Student’s _t_-test
Fig. 6
The metabolic switch in peritoneal tissue-resident macrophages is independent of nitric oxide and TLR signalling. a Line graph showing relative changes in oxygen consumption rate (OCR) vs atpenin A5 dose in peritoneal tissue-resident macrophages (pTRMØ). Data are from a single experiment (n = 5–6 separate wells per group), but the difference between basal and maximal OCR changes represents at least 3 experiments. b Line graph showing OCR vs time in pTRMØ. Cells were treated with or without zymosan (50 μg/ml, 2 h) before metabolic flux analysis; cell membranes were permeablised and rotenone (200 nM) added, succinate (20 mM) was added at the grey arrow, and atpenin A5 (1 μM) was added at the black arrow. Data represent at least two independent experiments and shows 6 separate wells per group. Data a, b were analysed by paired two-way ANOVA with Sidak’s post-tests. c Graph showing the increase in mitochondrial parameters from WT, MyD88−/− and Ticam1−/− pTRMØ calculated by comparing mock and zymosan treated samples as seen in Fig. 5b. Data show n = 6 separate wells per group from one experiment. d Graph showing OCR vs time, zymosan was added to wild-type (WT) and Tlr2−/− pTRMØ at the arrow indicated. Data show n = 5 separate wells per group from one experiment. e Graph showing OCR vs time, zymosan was added to wild-type (WT) and Nos2−/− pTRMØ at the arrow indicated. Data in the left panel show n = 5 separate wells per group pooled from two experiments. All error bars denote mean ± SEM
Fig. 7
The metabolic enhancement in peritoneal tissue-resident macrophages occurs independently of NOX2, but can be initiated with activation of protein kinase C. a Peritoneal tissue-resident macrophage (pTRMØ) uptake of pHrodo zymosan (50 μg/ml) from Ncf1 −/− and WT mice. Data show 3 separate wells per group and was not significantly different by Student’s _t_-test. b OCR changes in WT and Ncf1 −/− pTRMØ after addition of zymosan or phorbol–myristate–acetate (PMA, 1 μM) at the indicated arrow. The far-right line graph shows an expanded view of Ncf1 −/− data from the middle plot. Data from the WT show n = 3 separate wells per group from one experiment, whereas the Ncf1 −/− shows n = 5 per group from two separate experiments. Both are representative of at least three independent experiments. c The relative total oxygen consumption after addition of atpenin A5 or vehicle control in Ncf1 −/− pTRMØ. Data (n = 5 separate wells per group) represent two independent experiments and were analysed by Student’s _t_-test. d Graph showing change in OCR after addition of atpenin A5 or vehicle control. Post treatment (Post) is applied 90 min after zymosan addition, pre-treatment occurs before, in the absence of zymosan (Pre). Data (n = 10–12 separate wells per group) were pooled from two independent experiments and were analysed by two-way ANOVA (interaction p < 0.0001) with Sidak’s post-tests. e Quantification of the total oxygen consumption after addition of PMA in the presence of the indicated inhibitors (rotenone = 200 nM). Data (n = 3–6 separate wells per group) represent two independent experiments and were analysed by one-way ANOVA (p < 0.0001) with Tukey’s post-tests. f Graph showing change in OCR after addition of the indicated drugs. Post treatment (Post) is applied 40 min after PMA addition, pre-treatment occurs before, in the absence of PMA (Pre). Data (n = 5–6 separate wells per group) represent two independent experiments. Data were analysed by two-way ANOVA (interaction p < 0.0001) with indicated Sidak’s post-tests. g The protein kinase C inhibitor sotrastaurin (5 μM) or control were added to pTRMØ 30 min before addition of respiratory burst stimulants PMA or zymosan, data are quantified as relative total oxygen consumed post burst stimulants. Data (n = 5 separate wells per group) represent two independent experiments, analysed by two-way ANOVA (interaction p = 0.6854, sotrastaurin p < 0.0001) with indicated Sidak’s post-tests. All error bars denote mean ± SEM
Fig. 8
The electron transport chain is required for nominal reactive oxygen species production in peritoneal tissue-resident macrophages. a Bar graphs showing ATP and NADPH assays of peritoneal tissue-resident macrophages (pTRMØ). Cells were starved of glutamine (2 mM) and/or glucose (25 mM) 1 h before addition of zymosan (50 μg/ml) or control, extracts were taken after 2 h. Data were analysed by two-way ANOVA. For ATP, interaction p = 0.16, effects of zymosan or fuel p < 0.0001. For NADPH ratio, interaction p = 0.041. b Bar graphs showing ATP and NADPH assays of neutrophils. Cells were treated as in a, data were analysed by two-way ANOVA with Sidak’s post-tests for glutamine addition. For ATP, interaction p = 0.21, zymosan p = 0.15, fuel p < 0.0001. For NADPH ratio, interaction p = 0.85, zymosan p = 0.56, fuel p < 0.0001. c Bar graphs showing ATP and NADPH assays of pTRMØ. Cells were treated with inhibitors 30 min before addition of zymosan as in a. Oligomycin (1.26 μM), rotenone (200 nM), thenoyltrifluoroacetone (TTFA, 2 mM), antimycin A (1 μM), dehydroepiandrosterone (DHEA, 100 μM), 2-deoxyglucose (2-DG, 100 mM). Data were analysed by two-way ANOVA with Sidak’s post-tests vs control. For ATP, interaction p = 0.47, zymosan or drug were p < 0.0001. For NADPH ratio, interaction p = 0.036. Data a–c represent at least two independent experiments and show 3–4 separate wells per group. d Bar chart showing the median fluorescent intensity (MFI) of mitoSOX dye in Ncf1 −/− pTRMØ treated with phorbol–myristate–acetate (PMA, 1 μM) or control. Data (n = 4 separate wells per group) represent two independent experiments and were analysed by Student’s _t_-test. e Line graph (left) showing luminol luminescence against time in pTRMØ treated with zymosan 1 min before time 0. Bar chart (right) showing quantification of the area under the curve luminescence from the line graph in the presence of the indicated inhibitors (as in b + atpenin A5 (1 μM)). Data (n = 4–6 separate wells per group) were combined from two separate experiments, each treatment result is representative of at least two independent experiments. Data were analysed by one-way ANOVA with Tukey’s post-tests vs the matched control. All error bars denote mean ± SEM
Fig. 9
The mitochondrial electron transport chain is required for antimicrobial function in peritoneal tissue-resident macrophages. a Quantification of cytokines from peritoneal tissue-resident macrophage (pTRMØ) supernatant 24 h after Saccharomyces cerevisiae (S. cerevisiae) (50 μg/ml) ±dimethylmalonate (DMM,10 mM) or ±atpenin A5 (1 μM). The far-right panel shows IL-1β after 1 h post-stimulation with adenosine triphosphate (ATP, 3 mM), 23 h after S. cerevisiae. Data (n = 4 separate wells per group) represent two independent experiments. For TNF, IL-10 and KC, data were analysed by one-way ANOVA with Dunnett’s post-tests; for IL-1β, data were analysed by two-way ANOVA with Sidak’s post-tests vs vehicle control or vs no ATP as indicated (interaction p < 0.0001, ATP _p_ < 0.0001, complex-II _p_ < 0.0001). Error bars = mean ± SEM. b** Relative expression of Il1b RNA from samples in a. c Representative photographs of pTRMØ. In c, d, zymosan-AlexaFluor488 (5 μg/ml) are green, mitochondria labelled with Mitotracker Red CMX-Ros (25 nM) are red, whereas the nuclei are labelled blue with 6-diamidino-2-Phenylindole (DAPI,125 ng/ml). The yellow box denotes mitochondria surrounding zymosan. Zymosan cores can be seen with non-specific Mitotracker Red fluorescence. d A characteristic _z_-stack showing mitochondria surrounding the phagolysosome. The yellow numbers denote distance from coverslip. e Quantification of the number of Mitotracker Red positive pixels expressed as percentage of total (termed mitochondrial localisation %) at different radial distances from zymosan. Data shown are 53 zymosan particles from five images. Data were analysed by a one-way ANOVA (p < 0.0001) with a linear regression post-test shown. **f** Box and whisker plots showing quantification of size for >800 mitochondrial units per group taken from five images of pTRMØ in the presence or absence of zymosan-AlexaFluor488. g Representative pictures (left) of microbial plates with S. cerevisiae colonies. Data shown are from S.cerevisiae cultured ± pTRMØ with atpenin A5 (1 μM) or control for 4½ h. The graph on the right shows quantification of the average microbial killing. Data are from three experiments (n = 13 per group) and were analysed by two-way ANOVA (interaction + experimental repeat p = 0.26, atpenin A5 p = 0.10). h Representative pictures (left) and quantification of data (right) as in e, but with antimycin A (1 μM) or control. Data are from four experiments (n = 18 per group) and were analysed by two-way ANOVA (interaction + experimental repeat p = 0.10, antimycin A p < 0.0001). Whiskers e–h** show 10–90% of the range. White scale bars are 10 μm
Fig. 10
Peritoneal tissue-resident macrophages are metabolically poised to engage microbes using tissue-niche fuels Model showing the proposed metabolic phenotypes of resting- and phagocytic-peritoneal tissue-resident macrophages and how this impacts function. NAA, N-acetyl aspartate; PKC, Protein kinase C; Shuttles, malate-aspartate or citrate-malate mitochondrial shuttle systems
References
Publication types
MeSH terms
Substances
LinkOut - more resources
Full Text Sources
Other Literature Sources
Molecular Biology Databases