key: cord-0333188-rosit1pn authors: Woods, Parker S.; Kimmig, Lucas M.; Sun, Kaitlyn A.; Meliton, Angelo Y.; Shamaa, Obada R.; Tian, Yufeng; Cetin-Atalay, Rengül; Sharp, Willard W.; Hamanaka, Robert B.; Mutlu, Gökhan M. title: HIF-1α induces glycolytic reprogramming in tissue-resident alveolar macrophages to promote survival during acute lung injury date: 2022-03-01 journal: bioRxiv DOI: 10.1101/2022.02.28.482301 sha: b19cbe23dec3efb27bb8de07e09ec081b91fb5d6 doc_id: 333188 cord_uid: rosit1pn Cellular metabolism is a critical regulator of macrophage effector function. Tissue-resident alveolar macrophages (TR-AMs) inhabit a unique niche marked by high oxygen and low glucose. We have recently shown that in contrast to bone marrow-derived macrophages (BMDMs), TR-AMs cannot utilize glycolysis and predominantly rely on mitochondrial function for their effector response. It is not known how changes in local oxygen concentration that occur during conditions such as acute respiratory distress syndrome (ARDS) might affect TR-AM metabolism and function; however, ARDS is associated with progressive loss of TR-AMs, which correlates with the severity of disease and mortality. Here, we demonstrate that hypoxia robustly stabilizes HIF-1α in TR-AMs to promote a glycolytic phenotype. Hypoxia altered TR-AM metabolite signatures, cytokine production, and decreased their sensitivity to the inhibition of mitochondrial function. By contrast, hypoxia had minimal effects on BMDM metabolism. The effects of hypoxia on TR-AMs were mimicked by FG-4592, a HIF-1α stabilizer. Treatment with FG-4592 decreased TR-AM death and attenuated acute lung injury in mice. These findings reveal the importance of microenvironment in determining macrophage metabolic phenotype, and highlight the therapeutic potential in targeting cellular metabolism to improve outcomes in diseases characterized by acute inflammation. Glycolytic metabolism has been ascribed a central role in macrophage inflammatory processes 19 (Tannahill, Our group has recently demonstrated that TR-AMs rely predominantly on oxidative 36 phosphorylation under steady-state conditions and that glycolysis is dispensable for 37 proinflammatory effector function in these cells (Woods, Kimmig et al. 2020 ). Together, these 38 findings highlight the lung microenvironment's central role in dictating TR-AM responses. Conditions associated with severe airway inflammation (i.e., acute respiratory distress syndrome 41 (ARDS)) increase alveolar epithelial/endothelial barrier permeability (Ware and Matthay 2000) . To answer these questions, we used a variety of metabolic and immunological approaches. We 74 observed that HIF-1α was undetectable in primary TR-AMs cultured under normoxia, but was 75 robustly stabilized under hypoxia in a dose-dependent fashion. Upon hypoxic HIF-1α stabilization, 76 TR-AMs acquired a glycolytic phenotype, which was not observed under normoxic conditions. In Tissue-resident alveolar macrophages exhibit HIF-1α stabilization and develop a glycolytic 97 phenotype in response to hypoxia 98 We have recently shown that TR-AMs maintain a very low glycolytic rate which is not augmented 99 by activation of inflammatory responses (Woods, Kimmig et al. 2020 ). As TR-AMs inhabit an 100 environment with high oxygen levels, we hypothesized that TR-AMs may not be able to induce 101 glycolytic reprogramming in response to either inflammatory stimuli or to physiologic hypoxia. Glycolysis stress tests were performed following overnight (16 hours) exposure to decreasing 103 levels of ambient oxygen. Unlike the inability of TR-AMs to induce glycolysis after inflammatory 104 stimulus under normoxia, they exhibited a progressive increase in the extracellular acidification 105 rate (ECAR) in response to escalating degrees of ambient hypoxia ( Figure 1A ). Both basal rate 106 of glycolysis and glycolytic reserve increased substantially when oxygen levels were lowered to 107 3.0% and 1.5% ( Figure 1B ). HIF-1α levels were nearly undetectable under normoxic conditions; 108 however, with increasing degrees of hypoxia, HIF-1α stabilization occurred in a dose-dependent 109 fashion and was detectable in the nucleus ( Figure 1C ). Intracellular lactate levels increased 110 slightly in TR-AMs after exposure to hypoxia compared with normoxic control cells while the 111 lactate levels in the media increased by a greater degree ( Figure 1D ). Pretreating TR-AMs prior 112 to hypoxia with echinomycin, an inhibitor of HIF-1α DNA binding activity (Kong, Park et al. 2005 ), 113 disrupted hypoxia-induced increases in glycolytic output in a dose-dependent fashion ( Figure 1E ). Echinomycin also reduced hypoxia-induced increases in glycolytic protein expression (LDHA and 115 HK2), suggesting that HIF-1α is required for glycolytic adaption to hypoxia in TR-AMs ( Figure 1F ). Both short-term (2 hours) and prolonged (16 hours) exposure to hypoxia (1.5% O 2 ) led to 118 significant increases in nuclear HIF-1α protein levels in TR-AMs ( Figure S1A ). Glycolysis stress 119 tests demonstrated that short-term hypoxia treatment failed to induce significant alterations in 120 glycolysis or glycolytic capacity in TR-AMs compared to prolonged hypoxia treatment suggesting 6 that transcription and translation of glycolytic genes that are targets of HIF-1α are required ( Figure 122 S1B,C). Taken together, these data indicate that TR-AM HIF-1α stabilization in response to 123 hypoxia is dose-dependent, and that prolonged hypoxia, but not short-term hypoxia exposure, 124 leads to a functional glycolytic phenotype in TR-AMs. 2020). We found that, unlike TR-AMs, BMDMs exposed to hypoxia (16 hours) exhibit minimal 131 changes in glycolytic rate or glycolytic capacity (Figure 2A-B) . Interestingly, we found that BMDMs . To determine the effect of HIF-1α stabilization on TR-AM's effector response, we 170 measured the production of proinflammatory cytokines in response to LPS under hypoxia. TR-171 AMs were exposed overnight to hypoxia (1.5% O 2 ) or normoxia and then subsequently treated 172 with LPS while maintaining original O 2 conditions. Hypoxia alone did not stimulate cytokine 173 production without LPS treatment. Hypoxic TR-AMs secreted significantly higher levels of TNF-α, KC, and IL-1β in response to LPS compared to normoxic controls. In contrast, IL-6 and CCL2 We and others have shown that BMDMs exhibit an immediate enhancement in glycolytic output 186 in response to LPS ( Figure S4 ). It is thought that this increase in glycolysis following LPS supports 187 the proinflammatory response. We have shown that LPS-induced inflammation in TR-AMs is We have previously shown that unlike BMDMs, TR-AMs effector function is acutely sensitive to 206 mitochondrial inhibition (Woods, Kimmig et al. 2020 ). Since TR-AM capacity for glycolysis 207 expands with decreasing levels of O 2 , we next sought to assess mitochondrial function under 208 hypoxia and performed a mitochondrial stress test on TR-AMs that had been exposed to varying 209 oxygen concentrations. Interestingly, mild-to-moderate degrees of ambient hypoxia did not 210 appear to significantly alter overall mitochondrial function in these cells. Only 1.5% O 2 caused 211 significant reductions in oxygen consumption rate (OCR) across all mitochondrial parameters 212 ( Figure 5A , B). ECAR tracings during the mitochondrial stress test demonstrated that, other than 213 severe hypoxia (1.5% O 2 ), the majority of acid produced under mild-moderate hypoxia is derived 214 from CO 2 , as the application of rotenone and antimycin A led to a significant reduction in ECAR 215 ( Figure 5C ). When exposed to 1.5% O 2 , TR-AM energy is derived mostly from glycolysis with little 216 TCA activity and no significant contribution of CO 2 to extracellular acidification. Overall BMDM 217 mitochondrial function was impaired by 1.5% O 2 , but the effect was greatly diminished compared 218 to TR-AMs ( Figure S5A , B). BMDM ECAR tracing during mitochondrial stress test demonstrated 219 that most acid production remained unchanged in response to rotenone and antimycin A 220 regardless of O 2 concentration ( Figure S5C ). This suggests that BMDM acidification is 221 glycolytically-derived under both normoxia and hypoxia. TR-AM cytokine production in response to LPS was highly susceptible to inhibition by low doses 224 of ETC inhibitors, rotenone and antimycin A, under normoxic conditions. This effect was greatly 225 attenuated after exposure to hypoxia ( Figure 5D ). Additionally, high doses of ETC inhibitors 226 induce cytotoxicity in normoxic TR-AMs, but hypoxic preconditioning significantly enhanced TR-227 AM cell viability ( Figure 5E ). In contrast, BMDM cytokine production was only marginally affected 228 by ETC inhibition with the exception of observed decrease in IL-1β. Unlike TR-AMs, hypoxia did 229 not significantly alter BMDM cytokine production in the presence of ETC inhibitors ( Figure S5D ). Similarly, ETC inhibition did not to induce cytotoxicity in BMDMs under normoxia or hypoxia 231 ( Figure S5E ). Like hypoxia, FG-4592 treatment could also rescue ETC inhibitor-induced impairment in cytokine 282 production ( Figure 7H ) and cell death in TR-AMs ( Figure 7I ). However, unlike TR-AMs exposed hypothesized that failure to adapt to the hypoxic environment may play a role in the TR-AM loss. We found that HIF-1α activation was sufficient to promote glycolysis, and rescue TR-AM viability Gene counts were then imported into R for differential expression analysis using the Bioconductor 491 package DESeq2. Gene counts were filtered to remove low-expressing genes at a threshold of 2 Filtering of supernatant occurred at 9,100 x g at 4 °C for 2 hours. The filtrate was sent to HMT 516 and analyzed using capillary electrophoresis-mass spectrometry. In vitro cytotoxicity was measured using the SRB assay (Vichai and Kirtikara 2006 The data were analyzed in Prism 8 (GraphPad Software Inc.). All data are shown as mean ± SD. ANOVA was used for statistical analyses of data sets containing more than two groups, and Using Seahorse XF24 technology, glycolysis was measured as extracellular acidification rate (ECAR). (C) Interleaved scatter plots quantifying glycolytic parameters. Data represents at least 3 independent experiments (n=4 separate wells per group). Significance was determined by two-tailed Student's t test. *, p < 0.05. Figure S5 . The effect of hypoxia on BMDM mitochondrial function, cytokine production and cell liability under ETC inhibition. (A) Mitochondrial stress test to measure oxygen consumption rate (OCR) using Seahorse XF24 in BMDMs. (B) Interleaved scatter plots quantifying mitochondrial respiration parameters. Data represents at least 3 experiments (n=4 separate wells per group). Mitochondrial parameters were were compared against 21% O 2 and significance was determined by one-way ANOVA with Bonferroni's post test (C) ECAR measurement during mitochondrial stress test. (D) BMDMs were incubated overnight (16h) under 21% or 1.5% O 2 then stimulated with 20ng/ml LPS in the presence of absence of mitochondrial inhibitors (20nM Antimycin A (Ant) or Rotenone (Rot)) for 6 hours while maintaining pretreatment conditions. ELISA was used to measure secreted cytokine (TNFα, IL-6, KC, CCL2, and IL-1β) levels in media. ATP added to cells prior to collection for IL-1β assessment. Data represent at least 3 independent experiments; n=3 per group. Significance was determined by one-way ANOVA with Bonferroni's post test. (E) BMDMs were cultured under 21% or 1.5% O 2 for 6h then treated with mitochondrial inhibitors (100nM Ant or 500nM Rot) overnight and an Sulforhodamine B assay was performed to measure cytotoxicity. Graphs represent cell viability compared to control, 21% O 2 group. Data represent at least 3 independent experiments (n=3 per group). Significance was determined by two-way ANOVA with Bonferroni's post test. All error bars denote mean ± SD. *, p < 0.05. Regulation of alveolar macrophage death in acute lung 627 inflammation Metabolic 630 reprogramming of macrophages: glucose transporter 1 (GLUT1)-mediated glucose metabolism 631 drives a proinflammatory phenotype Hypoxia-Induced Inflammation in the Lung Depletion of Alveolar Macrophages 637 during Influenza Infection Facilitates Bacterial Superinfections Circuits between infected macrophages and T cells in SARS-CoV-2 pneumonia Downregulation of the antigen presenting cell function(s) of pulmonary dendritic cells in 668 vivo by resident alveolar macrophages Alveolar macrophages: plasticity in a tissue-specific context Anti-inflammatory 674 effect of IL-10 mediated by metabolic reprogramming of macrophages Macrophage Terminal Differentiation, Self-Renewal, and Function Modulation of Dendritic Cell Trafficking to and from the Airways NLRX1 689 prevents mitochondrial induced apoptosis and enhances macrophage antiviral immunity by 690 interacting with influenza virus PB1-F2 protein Real time discrimination of inflammatory macrophage 693 activation using Agilent Seahorse XF technology Alveolar macrophages are indispensable for controlling influenza viruses in lungs of 697 pigs Alveolar Macrophages Have a Protective Antiinflammatory Role during Murine 701 Echinomycin, a Small-Molecule Inhibitor of Hypoxia-706 Inducible Factor-1 DNA-Binding Activity Tissue-resident macrophage enhancer landscapes are shaped by the local 710 microenvironment Single-cell landscape of bronchoalveolar immune cells in 714 patients with COVID-19 Alveolar macrophage depletion increases the severity of acute inflammation following 718 nonlethal unilateral lung contusion in mice Myeloid HIF-1 Is Protective in