Get 20M+ Full-Text Papers For Less Than $1.50/day. Subscribe now for You or Your Team.

Learn More →

VSIG4 inhibits proinflammatory macrophage activation by reprogramming mitochondrial pyruvate metabolism

VSIG4 inhibits proinflammatory macrophage activation by reprogramming mitochondrial pyruvate... ARTICLE DOI: 10.1038/s41467-017-01327-4 OPEN VSIG4 inhibits proinflammatory macrophage activation by reprogramming mitochondrial pyruvate metabolism 1 1 1 1 1 1 2 2 Jialin Li , Bo Diao , Sheng Guo , Xiaoyong Huang , Chengying Yang , Zeqing Feng , Weiming Yan , Qin Ning , 3 1 1 Lixin Zheng , Yongwen Chen & Yuzhang Wu Exacerbation of macrophage-mediated inflammation contributes to pathogenesis of various inflammatory diseases, but the immunometabolic programs underlying regulation of mac- rophage activation are unclear. Here we show that V-set immunoglobulin-domain-containing 4 (VSIG4), a B7 family-related protein that is expressed by resting macrophages, inhibits −/− macrophage activation in response to lipopolysaccharide. Vsig4 mice are susceptible to high-fat diet-caused obesity and murine hepatitis virus strain-3 (MHV-3)-induced fulminant hepatitis due to excessive macrophage-dependent inflammation. VSIG4 activates the PI3K/ Akt–STAT3 pathway, leading to pyruvate dehydrogenase kinase-2 (PDK2) upregulation and subsequent phosphorylation of pyruvate dehydrogenase, which results in reduction in pyr- uvate/acetyl-CoA conversion, mitochondrial reactive oxygen species secretion, and macro- phage inhibition. Conversely, interruption of Vsig4 or Pdk2 promotes inflammation. Forced expression of Vsig4 in mice ameliorates MHV-3-induced viral fulminant hepatitis. These data show that VSIG4 negatively regulates macrophage activation by reprogramming mitochon- drial pyruvate metabolism. 1 2 Institute of Immunology, PLA, Third Military Medical University, Chongqing, 400038, China. Institute of Infectious Disease, Tongji Hospital of Tongji Medical College, Huazhong University of Science and Technology, Wuhan, 430030, China. Laboratory of Immunology, National Institute of Allergy and Infectious Diseases, NIH, Bethesda, Maryland, MD 20892, USA. Jialin Li and Bo Diao contributed equally to this work. Correspondence and requests for materials should be addressed to Y.C. (email: yongwench@163.com) or to Y.W. (email: wuyuzhang@tmmu.edu.cn) NATURE COMMUNICATIONS 8: 1322 DOI: 10.1038/s41467-017-01327-4 www.nature.com/naturecommunications 1 | | | ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-017-01327-4 acrophages are essential for innate immunity owing to that VSIG4 can deliver anti-inflammatory signals. Here, we show functions in host defense, tissue development and that VSIG4 antagonizes activation signals in macrophages by Mhomeostasis . Macrophage functional disparity is stimulating PI3K/Akt–STAT3 cascades, augmenting expression attributed to two distinct subgroups, namely M1 (classically of pyruvate dehydrogenase kinase-2 (PDK2), and inhibiting activated) and M2 (alternatively activated) macrophages .M1 pyruvate dehydrogenase (PDH) activity via phosphorylation. macrophages have been implicated in initiating and sustaining Therefore, VSIG4 restricts pyruvate metabolism in the mito- inflammation in response to INF-γ and/or lipopolysaccharide chondria during oxidative phosphorylation (OXPHOS), resulting (LPS), whereas IL-4 or IL-13 polarized M2 macrophages seem to in suppression of mtROS secretion and M1 differentiation. have immunoregulatory functions in parasitic infections, tissue Conversely, Vsig4 or Pdk2 deficiency enhances macrophage −/− inflammation, remodeling of damaged tissue, and tumor pro- activation. Vsig4 mice are more susceptible to HFD-induced gression . Uncontrolled M1 activation can cause tissue damage obesity in association with insulin resistance. These mice manifest and pathogenesis in inflammatory diseases, including athero- with more severe liver damage and mortality as a result of MHV- sclerosis, obesity, diabetes, rheumatoid arthritis and hepatitis .In 3 infection, probably owing to overloading of macrophage- a model of mouse obesity caused by high-fat diet (HFD), the mediated inflammation in vivo. Interestingly, forced over- accumulation of adipose tissue macrophages (ATMs) within the expression of Vsig4 ameliorates MHV-3-induced viral fulminant white adipose tissue (WAT) involves remodeling of the enlarged hepatitis. These data identify an inhibitory function of VSIG4 in WAT and induction of insulin resistance via secreting proin- macrophage-mediated inflammation. 5, 6 flammatory mediators IL-1β and TNF . Moreover, in a viral fulminant hepatitis model, murine hepatitis virus strain-3 (MHV- 3) infection induces a macrophage-dependent cytokine storm of Results −/− IL-1, TNF, TGF-β, leukotriene B4, and pro-coagulant fibrinogen- Vsig4 mice are more susceptible to HFD-induced obesity. like protein-2 (FGL2), which causes fibrin deposition in the liver Macrophages actively contribute to the pathogenesis of diet- 5, 6 and results in acute hepatic necrosis and lethality of susceptible induced obesity , making them seemingly a good model for mice . Therefore, attenuation of macrophage-mediated inflam- examination of the biological functions of VSIG4. For this, age- −/− mation is a plausible strategy for treating inflammatory disorders. matched C57BL/6 wild type (WT) and Vsig4 mice were fed a −/− The regulation of macrophage activation has been extensively HFD. Interestingly, we found that Vsig4 mice gained sig- studied, with evidence suggesting the involvement of multiple nificantly more body masses than their WT controls after 5 weeks intracellular signaling regulators, including membrane molecules, of HFD feeding (Fig. 1a). However, no significant differences of small noncoding RNAs, microRNAs and epigenetic-associated body masses between the two groups fed a normal chow diet mechanisms . Evidence from metabolic screening and microarray (NCD) (Supplementary Fig. 1a). Analytical microCT (μCT) analyses shows that LPS-activated macrophages have alterations imaging revealed an increase in fat mass throughout the bodies of −/− in mitochondrial metabolites, indicating that reprogramming of the Vsig4 obese mice, with substantially more abundant −/− mitochondrial metabolism may be involved in the regulation of visceral fat (Fig. 1b). Vsig4 obese mice also manifested with 9, 10 + macrophage activation . For example, NAD , as a mitochon- significant increases in abdominal wall fat and perirenal fat drial intermediate metabolite, can inhibit inflammation through compared to the WT controls. In parallel, we also observed a 11, 12 inactivating transcription factor NF-κB . Conversely, succi- substantial increase in the serum levels of triglyceride, cholesterol nate, another important mitochondrial metabolite, accumulates and free fatty acid, which correlated to increases in body fat in −/− −/− in LPS-activated macrophages and promotes the transcription of Vsig4 obese mice (Fig. 1c). Furthermore, Vsig4 obese mice Il1b by stabilizing hypoxia-inducible factor 1α (HIF-1α) . tended to develop with high liver triglyceride levels and steatosis Moreover, mitochondrial reactive oxygen species (mtROS) can (Fig. 1d), as well as enlarge adipocytes (Fig. 1e). These data imply activate the NLRP3 inflammasome and trigger bioprocessing of that Vsig4 deficiency renders mice susceptible to HFD-induced proinflammatory cytokines including pro-IL-1β and pro-IL-18 in obesity. 14, 15 5 macrophages . Therefore, inhibition of autophagy, which Obesity is often associated with insulin resistance , a fact that −/− impairs the removal of ROS-generating mitochondria and causes led us to postulate that Vsig4 mice might have a disturbed NLRP3 inflammasome activation and IL-1β secretion, attenuates glucose metabolism. We next examined the blood glucose levels 16 −/− glucose tolerance and insulin sensitivity . Additionally, mtROS of these obese mice. Vsig4 obese mice exhibited statistically can induce M1 activation through activating NF-κB and stabi- significant elevations in blood glucose relative to WT littermates 17, 18 lizing HIF-1α . Therefore, suppression of mtROS secretion after 15 h of fasting. Moreover, the 5-h-fasting insulin levels were can mitigate pathogenesis in alcoholic steatohepatitis and reduce also substantially elevated (Fig. 1f). An oral glucose tolerance test 19 −/− lethality in endotoxin-mediated fulminant hepatitis . Never- (GTT) illustrated that Vsig4 obese mice had significantly theless, the mechanisms underlying reprogramming of mito- higher serum levels of glucose and insulin in response to the chondrial metabolism during macrophage activation are unclear. glucose load compared to their WT littermate controls, suggesting V-set immunoglobulin-domain-containing 4 (VSIG4) is a a severely impaired glucose metabolism (Fig. 1g). Nevertheless, membrane protein belonging to complement receptor of the the blood glucose metabolism was similar between the two groups 20, 21 immunoglobulin superfamily (CRIg) . By binding comple- under NCD conditions (Supplementary Fig. 1b, c). The insulin ment component C3b, VSIG4 mediates clearance of C3b- tolerance test (ITT) also indicated the existence of a significantly −/− opsonized pathogens, such as Listeria monocytogenes and Sta- more resistance in Vsig4 obese mice compared to the WT phylococcus aureus . The expression of VSIG4 is restricted to controls (Fig. 1g). Additionally, as compared to WT controls, −/− tissue macrophages, including peritoneal macrophages and liver- western blot analyses revealed that Vsig4 obese mice had residential Kupffer cells. Moreover, VSIG4 marks a subset of diminished phosphorylation of IRS-1 (p-IRS-1) and phosphor- 22 ser473 macrophages that associates with diabetes resistance . VSIG4 ylation of Akt (p-Akt ) in the visceral adipose tissue (VAT), can functionally inhibit IL-2 production and T-cell proliferation muscle and liver tissues after administrated with insulin (Fig. 1h). by binding an unidentified T-cell ligand or receptor . Interest- These data indicate that Vsig4 deficiency promotes obesity- ingly, a VSIG4-Fc fusion protein seems to protect against devel- associated insulin resistance. opment of experimental arthritis , experimental autoimmune To better understand the mechanisms by which the VSIG4 24 25 uveoretinitis , and immune-mediated liver injuries , suggesting pathway prevents weight gain during HFD consumption, we 2 NATURE COMMUNICATIONS 8: 1322 DOI: 10.1038/s41467-017-01327-4 www.nature.com/naturecommunications | | | NATURE COMMUNICATIONS | DOI: 10.1038/s41467-017-01327-4 ARTICLE −/− addressed the food intake of these mice. There were no obvious Vsig4 mice exacerbate MHV-3-induced fulminant hepatitis. changes in food intake and stool output between the WT and We further investigated the potential role of Vsig4 in viral ful- −/− Vsig4 mice under HFD conditions (Supplementary Fig. 2), minant hepatitis caused by MHV-3 infection, in which the virus- implying that the differences in weight gain are due to reasons induced exaggerated inflammation causes severe pathogenesis other than food consumption. We therefore examined the status largely due to the M1 macrophage-dependent “cytokine storm” . −/− of the ATMs. Compared to WT littermates, flow cytometric assay Vsig4 and congenic C57BL/6 WT littermates were infected −/− −/− showed that the ATMs from Vsig4 obese mice expressed with MHV-3 (100 PFU/mouse). Noticeably, Vsig4 mice died higher amounts of proinflammatory factors (like pro-IL-1β, IFN- rapidly following MHV-3 infection in contrast to WT littermates γ, and TNF), which have been described to actively participate (log-rank test, p = 0.0339, Fig. 2a). H&E staining showed that the −/− into the pathogenesis of HFD-caused obesity and insulin infected Vsig4 animals had more severe liver necrosis and 26, 27 resistance (Fig. 1i). In parallel, dramatically higher levels of hepatocyte apoptosis at 48 h and 72 h of infection (Fig. 2b), along −/− these cytokines were accumulated in the VAT of Vsig4 mice, with significantly higher levels of serum alanine aminotransferase as detected by qRT-PCR (Fig. 1j) and western blot (Fig. 1k), (ALT) and aspartate aminotransferase (AST), the two liver respectively. These data suggest that Vsig4 deficiency initiates damage indicating enzymes released into the blood compared macrophage-mediated inflammation, which triggers HFD- with WT controls (Fig. 2c). Plaque assay data also showed that induced obesity and insulin resistance. Vsig4 deficiency promoted virus replication in liver tissues a bc –/– WT Vsig4 35 WT (n =5) * ** * –/– ** 5 * 2.5 *** 1.25 * 5 5 Vsig4 (n =5) **** ** ** 2.0 1.00 4 4 1.5 3 3 3 0.75 2 2 2 1.0 0.50 1 1 1 0.5 0.25 15 0 0 0 0 –/– WT Vsig4 Weeks of HFD feeding de –/– –/– WT Vsig4 WT Vsig4 5 *** * * 4 20 3 15 2 10 1 5 0 0 0 f h –/– –/– Vsig4 –/– Vsig4 WT 25 Vsig4 12 10 ** 6 –/– *** Vsig4 12 KD –+ + +–+ + + Insulin 5 8 * * 180 p-IRS-1 6 9 WT 6 473 3 60 pAKT 10 4 WT 2 WT 60 AKT 5 2 3 GTT GTT ITT 37 GAPDH 0 0 0 0 0 15306090120 0 15 30 45 60 0 15 30 45 60 180 p-IRS-1 Times (min) Times (min) Times (min) p-AKT 60 AKT i k GAPDH Il-1 Tnf IL-6 –/– WT Vsig4 p-IRS-1 p = 0.14 180 5 * 2.0 * 3 KD #1 #2 #3 #1 #2 #3 60 p-AKT 1.5 17 TNF AKT 3 60 25 IFN-γ 1.0 37 GAPDH 17 IL1-β 0.5 GAPDH Pro-IL-1β IFN-γ TNF 0 0 –/– WT Vsig4 –/– WT Vsig4 −/− −/− Fig. 1 Vsig4 mice are more susceptible to HFD-induced obesity with insulin resistance. Eight-week-old male Vsig4 mice and age-matched C57BL/6 WT controls were fed a HFD. a Body weight was measured and compared. The obese mice were sacrificed after 10 weeks of HFD feeding. b Fat distribution was detected by μCT. Yellow indicates subcutaneous fat and brown indicates that visceral fat. c Measurement of abdominal wall fat, perirenal fat, serum triglyceride, cholesterol, and free fatty acid. d Representative liver H&E staining (left), and intrahepatic triglyceride contents (right), scale bar = 20 μm, n = 10 per group. e Representative the architecture of adipose tissues stained by H&E (left), adipocyte size and cell numbers was calculated (right), scale bar = 20 μm, n = 10 per group. f The 15-h-fasting blood glucose levels and 5-h-fasting serum insulin levels. g GTT and ITT were performed in theses obese mice, ser473 n = 6 per group. h Western blot of the AKT, p-Akt , and p-IRS-1 in VAT, muscle and liver tissues of obese mice after 4 min of insulin administration, n = 4 per group. ATMs were isolated from obese mice. i Flow cytometry analyzing pro-IL-1β, IFN-γ, and TNF. Cytokines in VAT were detected by j qRT-PCR and k western blot. Error bar, s.e.m. *p < 0.05, **p < 0.01 and ***p < 0.001 (Student’s t-test). Data are representative of five (a) and three (c–i) independent experiments NATURE COMMUNICATIONS 8: 1322 DOI: 10.1038/s41467-017-01327-4 www.nature.com/naturecommunications 3 | | | WT –/– Vsig4 WT –/– Vsig4 WT –/– Vsig4 WT –/– Vsig4 WT –/– Vsig4 Count 15 h-fasting glucose (mM/l) Body mass (g) 5 h-fasting insulin (ng/ml) 7 mRNA relative (folds) Blood glucose (mM/l) Triglyceride (U/mL) mRNA relative (folds) Insulin (ng/ml) mRNA relative (folds) Abdominal wall fat (g) Blood glucose (mM/l) Perrienal fat (g) Cholesterol (U/ml) Adipocyte size (μm ) Liver Muscle VAT Triglyceride (U/ml) 5 2 Cell numb. (x10 )/cm Free fatty acid (U/ml) ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-017-01327-4 (Fig. 2d). This suggests that VSIG4 is capable of attenuating (Fig. 2e), western blot (Fig. 2g), and immunohistochemistry MHV-3-induced pathogenesis. (Supplementary Fig. 4). Therefore, dramatically higher levels of The macrophage-derived FGL2 and proinflammatory cyto- these proinflammatory cytokines were accumulated in the virus- −/− −/− kines including TNF, IL-1 and IL-6, play essential roles in the infected Vsig4 serum (Fig. 2h). Finally, Vsig4 mice 28–30 pathogenesis of MHV-3-induced fulminant hepatitis .We responded with severe fibrinogen formation, leading to increased therefore examined the expression of these mediators in liver coagulation and necrosis post infection (Supplementary peritoneal exudate macrophages (PEMs) and liver tissues. Fig. 4). VSIG4 has been described to be a negative regulator of T- Although Vsig4 deficiency did not seem to affect the basal cell activation , nevertheless, the secretion of proinflammatory expression level of these factors before viral infection (Supple- cytokines (TNF and IFN-γ), and the expression of activated + + mentary Fig. 3), qRT-PCR data showed that MHV-3 infection markers (CD25 and CD69) from CD4 as well as CD8 T cells in −/− −/− super-induced Vsig4 PEMs to express Fgl2, Tnf, Il-1β, and Il-6 72 h of MHV-3-infected Vsig4 livers was similar to their WT (Fig. 2e). Flow cytometric data also confirmed at protein levels controls (Supplementary Fig. 5). These results clearly demon- that inflammatory cytokines like pro-IL1-β, TNF, and IL-6, were strate that Vsig4 deficiency exacerbates macrophage-mediated −/− dramatically increased in virus-infected Vsig4 PEMs (Fig. 2f), inflammation, which deteriorates MHV-3 virus-induced FH. suggesting Vsig4 deficiency promotes macrophage-derived inflammation in vivo. Consistent with this, Vsig4 deficiency in VSIG4 attenuates LPS-induced macrophage activation in vitro. liver Kupffer cells also resulted in higher levels of these factors Our above data from in vivo experiments demonstrated that deposited in the infected liver tissues, as detected by qRT-PCR Vsig4 deficiency promotes macrophage-derived inflammation, we ab c d H&E staining TUNEL –/– –/– * Vsig4 WT Vsig4 WT ** 80 NS 60 0 0h 48h PI WT WT (n =5) * ** –/– 20 Vsig4 4000 1 (n =10) NS 0 123 4 0 300 –/– Vsig4 0 –/– Days post infection WT Vsig4 0 h 48 h PI –/– WT Vsig4 ef 8 2.0 6 0 h 12 h 0 h 12 h 0 h 12 h ** * * * 3 6 10.71 48.61 1.5 6.24 44.31 34.44 4.05 2 3 4 1.0 WT FSC 1 2 0.5 0 0 0 0 8.37 90.76 65.19 200 60 2.0 4.69 2.0 7.59 59.43 *** * * * –/– Vsig4 Live/death 150 45 1.5 1.5 100 1.0 1.0 F4/80 F4/80 F4/80 50 0.5 0.5 F4/80 0 0 0 0 –/– WT Vsig4 –/– –/– WT Vsig4 WT Vsig4 KD #1 #2 #3 #4 #1 #2 #3 #4 #1 #2 #3 #4 #1 #2 #3 #4 600 * 800 300 ** 4 NS 800 NS FGL2 3 600 400 NS 17 TNF * 31 Pro-IL-1β 300 2 400 400 NS NS * NS 200 NS 1 200 200 IL-1β p17 NS 21 IL-6 0 0 0 0 0 0 h 24 h 48 h 0 h 24 h 48 h 0 h 24 h 48 h 0 h 24 h 48 h 0 h 24 h 48 h IFN-γ –/– WT Vsig4 41.6 Actin-β 24 h PI 48 h PI −/− Fig. 2 Vsig4 deficiency exacerbates MHV-3-induced fulminant hepatitis. The Vsig4 mice and age-matched C57BL/6 WT littermates were infected with MHV-3 (100 PFU/mouse) via i.p. injection. a The survival was monitored. b H&E staining of liver, and TUNEL staining of cell apoptosis, scale bar = 20 μm, n = 5–8 per group, arrow indicated positive cells. c Serum ALT and AST levels at 0 h and 48 h post infection (PI), n = 5–8 per group. d Plaque assay of virus titers in livers at 48 h PI. e qRT-PCR analyzing proinflammatory cytokines in PEMs at 12 h and in liver tissues at 72 h of MHV-3 infection. f Flow cytometry analyzing TNF, pro-IL1-β, and IL-6 from PEMs after 12 h of virus infection. g Western blot analyzing proinflammatory cytokines in infected livers at 24 h and 48 h PI, n = 4 per group. h ELISA of serum concentration of proinflammatory mediators, n = 5–10 per group. Error bar, s.e.m. *p < 0.05, **p < 0.01, ***p < 0.001 and NS, p > 0.05. a was analyzed by log-rank test and others are calculated by Student’s t-test. Data are representative of six (a) and three (b–f, h) independent experiments 4 NATURE COMMUNICATIONS 8: 1322 DOI: 10.1038/s41467-017-01327-4 www.nature.com/naturecommunications | | | Livers % of survival PEMS Fgl2 mRNA relative Fgl2 mRNA relative Tnf mRNA relative Tnf mRNA relative 72 h 48 h 0 h Il-6 mRNA relative Il-6 mRNA relative Il-1 mRNA relative Il-1 mRNA relative FGL2 (pg/mL) SSC SSC SSC IFN-γ (ng/mL) Pro-IL-1β TNF (pg/mL) Serum AST (U/mL) Serum ALT (U/mL) TNF IL-1β (pg/mL) IL-6 IL-6 (pg/mL) Liver viral titers log (PFU/g) 10 NATURE COMMUNICATIONS | DOI: 10.1038/s41467-017-01327-4 ARTICLE then investigated the role of VSIG4 in regulating macrophage Previous work has shown that the complement C3b is the activation in response to LPS in vitro. PEMs that are isolated natural ligand of VSIG4 . To address whether VSIG4 regulates −/− from Vsig4 mice appeared to present with an abrupt surge of macrophage activation is dependent on C3b, we used lentiviral −/− M1-like proinflammatory gene transcripts, such as Il-1β and Tnf, vectors to overexpress VSIG4 in C3 BMDMs and subjected at the very early stage (1 h and 6 h) of LPS exposure (Fig. 3a). This these cells to LPS treatment (2 μg/ml). ELISA showed that the was validated by the existence of higher levels of IL-1β, TNF, and secretion of IL-6 and IL-1β was still dramatically downregulated −/− + −/− IL-6 protein in the supernatants of LPS-stimulated Vsig4 PEM in VSIG4 C3 BMDMs than their controls (Fig. 3h). These cultures compared to the WT counterparts (Fig. 3b). Western blot combined data suggest that VSIG4-mediated cytokine production also confirmed these results (Fig. 3c). In addition, surface in macrophages is C3b independent. expression of M1 activation markers, including B7-H1, B7-DC, −/− B7-H3, and CD40, was higher in LPS-treated Vsig4 PEMs (Fig. 3d). These results suggest that VSIG4 is important for VSIG4 reprograms pyruvate metabolism and mtROS genera- controlling macrophage activation. tion. To investigate the molecular mechanisms for VSIG4- To avoid the cellular heterogeneity of conventional PEMs, we mediated macrophage activation, we constructed a platform using next chose a macrophage line, RAW264.7 cells, as a homogeneous VSIG4 RAW264.7 cells to test the dependence of macrophage model to examine the functional specificity of VSIG4. RAW264.7 activation on VSIG4. Inspired by recent studies showing that cell cells are lack of Vsig4 transcription, but they with lentiviral- metabolism has an important function in supporting macrophage mediated restoration of Vsig4 expression (Len-Vsig4) exhibited a activation and polarization , we measured cell metabolism reduction in LPS-induced M1 gene (Il1b, Il6, and Tnf) transcripts thereafter. Figure 4a shows that VSIG4 did not affect LPS- compared to the control counterparts (Fig. 3e). Similarly, ELISA induced glucose uptake. However, it inhibited lactate, pyruvate data also showed that the levels of TNF, IL-6, IL-1β, and IL-12p40 and acelyl-CoA accumulation after 6 h of LPS administration. protein in the supernatants of the LPS-treated VSIG4 RAW264.7 The mitochondrial oxygen consumption rate (OCR) in the cultures were significantly lower than the controls (Fig. 3f). These macrophages was then investigated, although VSIG4 did not VSIG4 RAW264.7 cells were also incapable of mounting LPS- affect oxygen consumption under normal conditions (Supple- induced CD40 upregulation (Fig. 3g). These combined data imply mentary Fig. 6a, b), it appeared to drastically downregulate that VSIG4 inhibits LPS-induced macrophage activation in vitro. oxygen consumption after LPS exposure, both in basal and ab c –/– WT Vsig4 Il-1 Tnf 400 20 NS NS 400 KD 015 45 60 0 15 45 60 LPS (min) 4 12 NS pro-IL-1β 300 15 3 9 NS IL1β p17 * 17 200 10 NS 200 2 6 NS TNF NS NS * 3 100 100 1 NS NS NS IL-6 0 0 0 0 41.6 Actin-β 016 (h) 016 (h) 016 24 (h) 01624 (h) 016 24(h) –/– –/– WT Vsig4 WT Vsig4 de f NS 6 30 Il-1 Il-6 Tnf *** * 4 20 20 4 8 * * 2 NS 10 NS NS 15 3 10 2 0 1 624 (h) 0 12 6 4 (h) 4 Len-Cont. PD-L1 PD-L2 NS Len-Vsig4 NS 2 5 1 NS NS ** NS 150 60 0 0 ** NS 0 6 (h) 06 (h) 0 6 (h) 100 40 NS NS NS Len-cont. Len-Vsig4 50 20 B7-H3 CD40 0 0 –/– 016 24(h) 0 1 6( 24h) WT Vsig4 8000 40 0 h 6 h LPS (2μg/ml) * NS Len-Cont. p = 0.07 Len-cont. 4000 * Len-Vsig4 NS Len-Vsig4 NS NS 026 12(h) 0 6 12 (h) CD40 Fig. 3 VSIG4 impedes LPS-induced macrophage M1 polarization in vitro. PEMs were treated with LPS (2 μg/ml), a qRT-PCR analysis of Il-1β and Tnf transcripts. b ELISA of cytokines in cultured supernatants. c Western blot analyzing cytokine protein expression. d Flow cytometry analyzing surface expression of activation markers. RAW264.7 cells stably infected with lentiviral control vectors (Len-cont.) or vectors encoding Vsig4 (Len-Vsig4), cells were further treated with LPS (2 μg/ml), e qRT-PCR analysis of Il-1β, Il-6, and Tnf transcripts. f ELISA detecting cytokines in cultured supernatant. g Flow −/− cytometry analyzing surface expression of CD40. h C3 BMDMs were tranfected to overexpress VSIG4, and cells were further treated with LPS (2 μg/ ml), the secretion of IL-6 and IL-1β was detected by ELISA. Error bar, s.e.m. *p < 0.05, **p < 0.01, ***p < 0.001 and NS, p > 0.05 (Student’s t-test). Data are representative of three independent experiments NATURE COMMUNICATIONS 8: 1322 DOI: 10.1038/s41467-017-01327-4 www.nature.com/naturecommunications 5 | | | Count Count Count mRNA relative (folds) mRNA relative (folds) mRNA relative (folds) Il-1β (pg/ml) mRNA relative (folds) IL-6 (ng/ml) TNF (ng/ml) mRNA relative (folds) Il-6 (pg/ml) IL-1β (pg/ml) IL-1β (pg/ml) TNF (ng/ml) IL-12 p40 (ng/ml) IL-6 (ng/ml) ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-017-01327-4 maximal OCR (Fig. 4b, c), implying that VSIG4 inhibits mito- a substrate that in turn stimulates HIF-1α-dependent IL-1β chondrial oxidation during macrophage activation. expression . Nevertheless, LPS-induced succinate accumulation It is believed that glucose oxidation via the mitochondrial and HIF-1α upregulation in macrophages appear to be unaffected electron transport chain is a major source of mtROS upon cells by the presence of VSIG4 (Supplementary Fig. 7), suggesting undergoing aerobic metabolism . The fact that VSIG4 inhibited VSIG4-mediated macrophage activation is HIF-1α independent. mitochondrial oxidation led us to investigate the status of mtROS, which can induce macrophage M1 activation through activating NF-κB and stabilizing HIF-1α . Compared to the controls, flow VSIG4 enhances PDK2 expression in macrophages. The cytometry showed that VSIG4 RAW264.7 cells had a signifi- observation of VSIG4 inhibiting mitochondrial oxidation and cantly less mtROS secretion, especially in response to LPS mtROS secretion led us to investigate the underlying molecular stimulation (Fig. 4d). Conversely, LPS exposure caused increased mechanisms. Regulation of pyruvate metabolism largely relies on −/− mtROS secretion in Vsig4 PEMs in vitro (Fig. 4e). Moreover, PDH, whose activity is inhibited by PDKs via phosphorylation . −/− both ATMs from Vsig4 obese mice and PEMs isolated from In examination of the 4 Pdk isoform expressions in BMDMs and −/− MHV-3-infected Vsig4 mice had higher levels of mtROS than we found that the Pdk2 mRNA and protein levels were sig- −/− their WT littermates (Fig. 4f). Inhibition of mtROS production by nificantly lower in Vsig4 macrophages than that in their WT using diphenyliodonium chloride (DPI) can efficiently block LPS- counterparts (Fig. 5a, b). This low level of PDK2 was responsible induced IL-6 secretion from both RAW246.7 and VSIG4 for the appreciable decreases in phosphorylation of PDH (p- S300 S293 RAW264.7 cells (Fig. 4g). This implies that VSIG4 inhibits PDH-E1α and p-PDH-E1α ), as detected by western blot macrophage M1 activation mainly by reducing pyruvate oxida- and immunofluoresence staining (Fig. 5b, c). The absence of tion and mtROS generation. Vsig4 in Kupffer cells also resulted in PDK2 reduction and lower S300 S293 The other important consequence of LPS-induced metabolic p-PDH-E1α /E1α levels in the liver tissues, both in unin- reprogramming in macrophages is the accumulation of succinate, fected and at 48 h of MHV-3-infected conditions (Fig. 5d). 8 60 200 NS 6 150 12 NS 8 4 30 NS 100 Len-Cont. NS Len-Vsig4 2 15 50 4 0 0 0 0 bc d Oligo FCCP AA+AR Basal OCR Maximal OCR LPS (2μg/mL) Mock * * 120 160 * * Len-cont. 100 Len-Vsig4 90 120 60 80 015 35 55 75 0 0 mtROS Times (min) Isotype Len-cont. Len-Vsig4 Len-cont. Len-Vsig4 ef 0 h 6 h LPS (2 μg/ml) Obese MHV-3 12 h PI Mock LPS LPS+DPI mtROS mtROS Len-cont. Len-Vsig4 –/– –/– WT Vsig4 WT Vsig4 Fig. 4 VSIG4 reprograms pyruvate metabolism and mtROS generation. VSIG4 RAW264.7 cells and their controls were treated with LPS (2 μg/ml) for 6 h. a Colorimetric/Fluorometric assay of glucose uptake, pyruvate, Acelyl-CoA and Lactate levels, n = 5 per group. b OCR of LPS-treated RAW264.7 cells by Seahorse XFp assay. OCR detected before and after sequential treatment with ATP synthase inhibitor Oligo, mitochondrial uncoupling agent FCCP, ETC inhibitors AA+AR at indicated times, n = 5 per group. c OCR at basal and maximal levels of the indicated conditions was plotted in bar graphs. d mtROS −/− secretion was detected by flow cytometry. e Vsig4 PEMs and the WT controls were treated with LPS for 0 and 6 h, and mtROS secretion was detected by flow cytometry. f mtROS secretion from ATMs of obese mice and PEMs from 12 h of MHV-3-infected animals was compared by flow cytometry. g RAW264.7 cells were treated with mtROS inhibitor DPI (10 μM) for 48 h in advance, cells were then added with LPS (2 μg/ml) for an additional 6 h, IL-6 in the supernatant was detected by ELISA. Error bar, s.e.m. *p < 0.05 and NS, p > 0.05 (Student’s t-test). Data are representative of three independent experiments 6 NATURE COMMUNICATIONS 8: 1322 DOI: 10.1038/s41467-017-01327-4 www.nature.com/naturecommunications | | | Untreated LPS (2 μg/ml) Untreated LPS (2 μg/ml) Untreated LPS (2 μg/ml) Untreated LPS (2 μg/ml) Count OCR pmol/min Glucose uptake μg/μg protein Pyruvate (ng/μL) Count OCR pmol/min Acelyl-CoA nmol/μg protein OCR pmol/min Lactate (RU) Count IL-6 (ng/ml) NATURE COMMUNICATIONS | DOI: 10.1038/s41467-017-01327-4 ARTICLE Conversely, transient expression of VSIG4 in RAW264.7 cells with a significant reduction in IL-6/TNF secretion and limitation appeared to enhance PDK2 expression, and increase PDH- of CD40 upregulation (Fig. 5i, j). Collectively, these results S300 E1α phosphorylation, especially after LPS administration suggest that VSIG4 controls macrophage M1 activation by (Fig. 5e). Therefore, the PDH activity was decreased (Fig. 5f). regulating the PDK2-dependent pyruvate mitochondria meta- These data suggest that VSIG4 promotes PDK2 upregulation in bolic axis. macrophages. To validate the role of PDK2 in LPS-induced macrophage activation, Pdk2 expression was silenced in RAW264.7 cells by VSIG4 promotes DK2 by activating PI3K/Akt–STAT3.To using the specific shRNA (sh-Pdk2). Interestingly, knock-down of address the underlying mechanism of VSIG4 promotes PDK2 Pdk2 appeared to enhance oxygen consumption, both under upregulation in macrophages, we focused on the PI3K–Akt sig- normal conditions (Supplementary Fig. 6c, d) and after 2 h of LPS naling machinery in consideration of the fact that this pathway is administration (Fig. 5g), implying that PDK2 inhibits mitochon- essential for cellular metabolism, in addition to other functions drial oxidation. Moreover, enhancing mitochondrial oxidation such as cell growth, survival etc. . Western blotting data showed −/− caused the elevation of the basal mtROS production, especially that Vsig4 PEMs and BMDMs had obviously decreased in Akt after LPS stimulation (Fig. 5h), together with augmenting IL-6 phosphorylation (Fig. 6a). However, lentivirus-mediated over- and TNF secretion (Fig. 5i), as well as promoting CD40 expression of VSIG4 in RAW264.7 cells resulted in higher p- −/− ser473 expression (Fig. 5j). Similarly, LPS-treated Pdk2 BMDMs also Akt expression compared to mock-infected controls manifested with enhanced mtROS secretion, promoted the (Fig. 6b), suggesting that VSIG4 transfers a feedback signal, transcription of proinflammatory cytokine genes, as well as licensing macrophages for Akt activation. To identify the func- enhanced CD40 expression compared to WT counterparts tional motifs of VSIG4 that are responsible for augmenting Akt (Supplementary Fig. 8). Conversely, lentiviral overexpression of phosphorylation, we created a series of point and truncation PDK2 in RAW264.7 cells has opposite effects, not only by mutants of the molecule and tested their activity in affecting Akt ser473 quenching of the basal level but also by preventing the LPS- activation. The data from probing with p-Akt suggested that induced upregulation of mtROS (Fig. 5h). This was associated the c-terminal residues 267–280 aa of VSIG4 were critical for ac d e –/– WT WT Vsig4 –/– Vsig4 –/– KD WT Vsig4 NS 2 46 PDK2 NS NS #1 #2 #3 #4 #1 #2 #3 #4 KD S300 43 p-PDH-E1α 1 46 PDK2 s293 Actin-β 43 p-PDH-E1α 41.6 S300 43 p-PDH-E1α 015 45 LPS (min) 43 PDH b f 37 GAPDH s293 KD p-PDH-E1α 46 PDK2 S300 43 p-PDH-E1α 200 NS S300 43 p-PDH-E1α 43 PDH s293 43 p-PDH-E1α 37 GAPDH 43 PDH S300 p-PDEH-E1α /Mito tracker/DAPI 0 41.6 Actin-β 0 h 6 h (LPS) PEMs BMDMs Len-Cont. Len-Vsig4 hi j NS NS 0 h 6 h LPS (2 μg/ml) 150 Oligo FCCP AA+AR 1500 5 * * * * 1200 4 Cont. ** Sh Pdk2 900 3 sh-Scr 600 sh-Pdk2 NS NS 300 1 2 3 4 2 3 4 010 10 10 010 10 10 0 h 6 h 12 h (LPS) 0 h 6 h 12 h (LPS) 0 mtROS Red CD40 0 255075 100 sh-Scr sh-Pdk2 sh-Pdk2 sh-Scr Times (min) Basal OCR Maximal OCR *** 12 80 *** Len-Cont. 120 160 * * Len-Pdk2 *** *** NS 60 80 NS 3 20 2 3 4 2 3 4 010 10 10 010 10 10 30 40 0 CD40 mtROS Red 0 h 6 h 12 h (LPS) 0 h 6 h 12 h (LPS) 0 0 Len-Cont. sh-Pdk2 Len-Cont. Len-Pdk2 Len-Cont. Pdk2-shRNA −/− Fig. 5 VSIG4 triggers PDK2 expression in macrophages. Macrophages from WT and Vsig4 mice were collected. a qRT-PCR detection of 4 Pdk isoforms s300 s293 s300 in BMDMs. b Western blot analyzing PDK2, p-PDH-E1α , p-PDH-E1α , and total PDH. c The location of p-PDH-E1α in mitochondria was analyzed s300 s293 by immunofluoresence double staining, scale bar = 20μm. d Western blot of PDK2, p-PDH-E1α , p-PDH-E1α in liver tissues at 0 h and 48 h PI. RAW264.7 cells were transfected to expression of Vsig4, and cells were further treated with LPS (2 μg/ml), e Western blot analysis of PDK2 and p-PDH- s300 E1α . f PDH activity analysis, n = 6 per group. The expression of Pdk2 in RAW264.7 cells was silenced by shRNA or enhancing Pdk2 expression by lentivirus infection. g Seahorse analysis of OCR after 2 h of LPS treatment (up), and basal and maximal OCR of the indicated conditions was plotted in bar graphs (down), n = 5 per group. h Flow cytometric assay of mtROS secretion after LPS administration. i ELISA of IL-6 and TNF in cultured supernatants, n = 4 per group. j Flow cytometric assay of LPS-caused CD40 expression at 6 h. Error bar, s.e.m. *p < 0.05,**p < 0.01, ***p < 0.001 and NS, p > 0.05 (Student’s t-test). Data are representative of three independent experiments NATURE COMMUNICATIONS 8: 1322 DOI: 10.1038/s41467-017-01327-4 www.nature.com/naturecommunications 7 | | | Pdk1 Pdk2 WT –/– Pdk3 Vsig4 Pdk4 WT –/– Vsig4 Len-Cont. Len-Vsig4 Len-Cont. Len-Vsig4 Len-Cont. Len-Vsig4 OCR pmol/min OCR pmol/min mRNA relative (folds) OCR pmol/min PEMs BMDMs Count Count 48 h 0 h IL-6 (pg/ml) IL-6 (ng/ml) TNF (ng/ml) TNF (ng/ml) PDH activity Count Count ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-017-01327-4 mediating Akt phosphorylation (Fig. 6b). Several families of resulting in PDK2 downregulation (Fig. 6g). Conversely, silenced kinases phosphorylate both serine and threonine residues in Stat3 expression in VSIG4 RAW264.7 cells by specific shRNA target substrates, thus result in three dimensional changes of the also decreased basal PDK2 expression (Fig. 6h). Furthermore, protein structure and thereby alter its enzymatic activity or affects using ChIP-qPCR, we found that LPS induces p-STAT3 its ability to interact with other proteins . There are two serine recruitment to one of the two putative binding sites at the 273 276 270 residues (Ser and Ser ) and two threonine residues (Thr −1,298 bp but not the −2,934 bp of Pdk2 promoter region, and and Thr ) in the c-terminal 267–280 residues of VSIG4, and the present of VSIG4 signaling markedly promotes this recruit- 270 273 further assessment revealed that mutation of Thr , Ser , and ment (Fig. 6i). Taken together, these data show that VSIG4 276 274 Ser , but not Thr , to Ala respectively, could successfully induces PDK2 expression via activating the PI3K/ Ser473 inhibited LPS-caused Akt phosphorylation in RAW264.7 Akt–STAT3 signaling pathway. cells, thus prevented PDK2 upregulation (Fig. 6c), suggesting To address whether VSIG4 promotes PDK2 expression in 270 273 276 Thr , Ser , and Ser residues of VSIG4 play essential role in macrophage is dependent on C3b, the human monocyte cell line, mediating PDK2 expression. The direct involvement of Akt in THP-1 cells, was transfected to overexpress human VSIG4, and modulating PDK2 expression was demonstrated by showing that these VSIG4 THP-1 cells were further induced to be macro- treating VSIG4 Raw264.7 cells with the Akt inhibitor MK-2206 phages by PMA stimulation. Additionally, cells were activated (Fig. 6d) or the PI3K inhibitor Ly294002 (Fig. 6e), all resulted in with microbeads-conjugated C3b, and western blot indicates that typical downregulation of p-Akt and PDK2. These combined data microbeads-C3b does not affect the basal and LPS-induced PDK2 −/− suggest that VSIG4 triggers PDK2 upregulation by activating the expression (Fig. 6j). Furthermore, overexpress VSIG4 in C3 PI3K/Akt pathway. BMDMs also increased basal and LPS-induced PDK2 expression Chromatin immunoprecipitation and massive parallel sequen- (Fig. 6j). These combined data suggest that VSIG4-mediated cing (ChIP-Seq) have demonstrated that the Pdk2 promoter PDK2 upregulation in macrophages is C3b independent. region has two binding sites for the signal transduction and activator of transcription-3 (STAT3) , which provides a basis for analyzing the signaling pathways responsible for VSIG4-PI3K/ Promoter methylation inhibits Vsig4 gene transcription. Our Akt-induced PDK2 upregulation. Interestingly, we found that data indicate that VSIG4 suppresses macrophage-dependent both the PI3K inhibitor Ly294002 (Fig. 6e), and the Akt inhibitor inflammation by augmenting PDK2 expression, which high- MK-2206 (Fig. 6f), could successfully inhibit LPS-induced STAT3 lights a plausible therapeutic intervention for inflammatory dis- phosphorylation (p-STAT3) in VSIG4 RAW264.7 cells. Simi- orders through enhancing VSIG4 signaling. However, in larly, the STAT3 inhibitor S3I-201 was also able to impair LPS- agreement with the previous report , we found that the PEMs induced p-STAT3 expression in VSIG4 RAW264.7 cells, thus and liver tissues isolated from MHV-3-infected mice manifested ab c SP IgV TM CT KD KD ser473 60 p-Akt ser473 Vsig4 60 p-Akt KD Δ119–280 ser473 46 PDK2 60 p-Akt 60 Akt Δ257–280 41.6 Actin-β 41.6 Actin-β 41.6 Actin-β Δ267–280 PEMS BMDMs 0h 1h LPS (2 μg/ml) – + + + + + LPS (2 μg/ml) de f – – – + LPS (2 μg/ml) – + + ++ LPS (2 μg/ml) – – + + LPS (2 μg/ml) – + – +++ LPS (2 μg/ml) – + – + Ly294002 (10 μM) KD 00 1 5 10 MK-2206 (μM) KD KD – + – + S3I-201 (100 μM) – 0.5 MK-2206 (μM) ser473 – 15 10 ser473 80 p-Akt KD 60 p-Akt 80 p-STAT3 80 STAT3 Ser307 80 p-STAT3 60 Akt STAT3 80 p-STAT3 46 PDK2 46 PDK2 46 PDK2 41.6 Actin-β Actin-β 41.6 Actin-β 41.6 41.6 Actin-β hi –/– C3 BMDMs THP-1 cells 1.5 * ––– +++ C3b (20 μg/ml) Len-cont.–RAW264.7 mouse IgG *** KD 1.0 0 h 1 h 3 h 0 h1 h3 h LPS (2 μg/ml) KD Len-cont.–RAW264.7 p-STAT3 mAb KD 46 PDK2 * NS + 46 PDK2 46 PDK2 Vsig4 RAW264.7 mouse IgG 0.5 Actin-β NS NS + 41.6 41.6 Actin-β 41.6 Actin-β Vsig4 RAW264.7 p-STAT3 mAb – – + + LPS (2 μg/ml) –1298 –2934 ser473 Fig. 6 VSIG4 promotes PDK2 expression through activating PI3K/Akt–STAT3. a Western blot of Akt and p-Akt expression. b RAW264.7 cells were ser473 infected with different Vsig4 deletion constructs, cells were further treated with LPS (2 μg/ml), the expression of p-Akt was analyzed by western blot. ser473 + 273 276 270 274 + c Western blot of p-Akt and PDK2 in VSIG4 RAW264.7 cells with mutation of Ser , Ser , Thr , and Thr to Ala. VSIG4 RAW264.7 cells ser473 were treated with d the Akt inhibitor MK-2206, e the PI3K inhibitor Ly294002, and the expression of Akt, p-Akt , PDK2, and p-STAT3 was detected by western blot. f VSIG4 RAW264.7 cells were treated with the MK-2206, and the expression of STAT3/p-STAT3 was analyzed by western blot. g Western blot of p-STAT3/STAT3 and PDK2 in LPS-activated VSIG4 RAW264.7 cells followed with STAT3 inhibitor, S3I-201 (100 μM) treatment for 24 h. h Western blot of PDK2 in Stat3 silenced VSIG4 RAW264.7 cells. i The enrichment of p-STAT3 in Pdk2 gene promoter region was detected by ChIP-qPCR assay. j Human VSIG4 THP-1 cells were treated with microbeads-C3b (20 μg/ml) in the presence of LPS (2 μg/ml), and the expression of PDK2 was −/− detected by western blot. Moreover, C3 BMDMs were transfected to overexpress VSIG4, and cells were further treated with LPS (2 μg/ml) for an additional 3 h, and the expression of PDK2 was detected by western blot. Error bar, s.e.m. *p < 0.05, ***p < 0.0001 and NS, p > 0.05 (Student’s t-test). Data are representative of three independent experiments 8 NATURE COMMUNICATIONS 8: 1322 DOI: 10.1038/s41467-017-01327-4 www.nature.com/naturecommunications | | | WT –/– Vsig4 WT –/– Vsig4 Sh-Scr Sh-Stat3 #1 Sh-Stat3 #2 Mock Vsig4Δ119–280 Vsig4Δ257–280 Vsig4Δ267–280 Vsig4 Mock Vsig4Δ119–280 Vsig4Δ257–280 Vsig4Δ267–280 Vsig4 Vsig4 Thr →Ala Ser →Ala Thr →Ala Ser →Ala Vsig4 Len-Cont. Len-Vsig4 Len-Cont. Len-Vsig4 Binding (%) NATURE COMMUNICATIONS | DOI: 10.1038/s41467-017-01327-4 ARTICLE with lower VSIG4 expression as compared to the uninfected negatively regulates Vsig4 gene transcription (Fig. 7f). In controls (Fig. 7a, b). Moreover, administration of proin- accordance, proinflammatory stimuli seemed to have no addi- flammatory mediators (including LPS, TNF, MALP2, IFN-γ, poly tional inhibition on the promoter activity in RAW264.7 cells (I:C) or CpG) apparently can induce a transient sharp decline of transfected with M.SssI-methylated constructs (Fig. 7f). These Vsig4 gene transcription and protein in ex vivo PEMs (Fig. 7c, d). data demonstrate that Dnmt3a controls Vsig4 gene repression Thus, rapid VSIG4 downregulation appears to be a common through fast methylation of Vsig4 gene promoter. response of macrophages upon inflammatory stimulations. We also analyzed the genomic DNA sequences of isolated The epigenetic mechanisms, especially DNA methylation of BMDMs that were treated with various proinflammatory CpG sites within promoter regions, have recently been described mediators for 12 h. Proinflammatory stimuli all appeared to to mediate gene silencing . There are three types of mammalian induce very high incidence (98–100%) of methylation at a CpG DNA methyltransferases (Dnmt), distinct from Dnmt1 that is site (-374 bp) in the promoter region of Vsig4 gene, which is responsible for copying DNA methylation patterns during significantly elevated from a 75% basal methylation at this site in replication, Dnmt3a and Dnmt3b are important in de novo the untreated controls as detected by using the Sequenom DNA methylation . We thereafter hypothesized that inflamma- MassARRAY platform (Supplementary Fig. 10a, b, Supplemen- tion leads to silencing Vsig4 gene transcription through triggering tary Table 1). Furthermore, ChIP-qPCR assays reveal a significant the transcriptional activation of Dnmts. Western blot showed that enrichment of Dnmt3a in −374 bp of Vsig4 promoter region after the expression of Dnmt3a was upregulated in BMDMs in the BMDMs were treated with proinflammatory factors (Supple- response to proinflammatory stimuli (Fig. 7e), nevertheless, the mentary Fig. 10c). These combined data suggest that CpG at -374 expression of Dnmt1 and Dnmt3b appeared to be not affected bp site, probably with other CpG islands in the Vsig4 promoter under such conditions (Supplementary Fig. 9). However, 5-aza- region, play an essential role for negative feedback control of 2ʹ-deoxycytidine (AZAdC), a general DNA methyltransferases macrophage activation during inflammatory response. inhibitor, was able to effectively downregulate Dnmt3a expres- sion, and in agreement, macrophages treated with AZAdC appeared to be resistant to the proinflammatory factors-caused Forced overexpression of Vsig4 improves MHV-3-caused FH. VSIG4 downregulation (Fig. 7e). We therefore tried to transiently force the expression of Vsig4 in To validate the importance of methylation in controlling Vsig4 MHV-3-susceptible C57BL/6 WT mice using lentiviral vectors gene transcription, a 840 bp fragment of Vsig4 gene promoter in vivo. These mice expressed significantly higher levels of VSIG4 (−840/ + 1) was constructed into a luciferase reporter pGL3-basic in the livers on day 6 of transduction compared to control vector. The pGL3-Basic and Vsig4 promoter constructs (−840/ + infected animals (Fig. 8a). Interestingly, Vsig4-transgenetic mice 1) were further fully methylated by CpG Methyltransferase (M. had significant enhancing PDK2 expression while lessening PDH S300 SssI). We transfected RAW264.7 cells with these plasmids and phosphorylation (p-PDH-E1α ) in liver tissues at 72 h of compared their reporter luciferase activities. Interestingly, the MHV-3 infection (Fig. 8b), leading to lower levels of FGL2, TNF, luciferase activity of the M.SssI-methylated construct exhibited IL-1β, and IL-6 deposition in liver tissues (Fig. 8b), together with 53% reduction in promoter activity, indicating DNA methylation reducing liver damage (Fig. 8c), along with a considerably ab c d PBS TNF 1.5 Poly(I:C) LPS 12 h PI 0 h 48 h PI IFNγ LPS+IFN-γ MALP-2 CpG KD KD #1 #2 #1 #2 1.0 0 h PI 45 VSIG4 45 VSIG4 41.6 Actin-β 41.6 Actin-β 0.5 0 1 2 3 4 ** 10 10 10 10 10 ** VSIG4 ef pGL3-basic pGL3-Vsig4 promoter (–860/+1) 1.5 NS NS NS * Unmethylated KD NS 1.0 M.Sssl methylated NS 101.5 Dnmt3a NS NS 45 VSIG4 0.5 41.6 Actin-β Fig. 7 Vsig4 gene transcription is repressed Dnmt3a-mediated DNA methylation. The C57BL/6 WT mice were infected with MHV-3 (100 PFU/mouse), a the expression of VSIG4 on PEMs at 0 h and 12 h PI was detected by flow cytometry. b VSIG4 protein level in liver tissues was analyzed by western blot. The BMDMs were treated with proinflammatory factors for 12 h. c Vsig4 gene transcription was detected by qRT-PCR. d VSIG4 protein expression was evaluated by western blot. e The BMDMs were treated with Dnmts inhibitor-AZAdC (10 μM) for 72 h in advance, cells were then further added with proinflammatory mediators for 12 h, the expression of Dnmt3a and VSIG4 was assessed by western blotting. f Luciferase activity of the lysates from RAW264.7 cells transfected with unmethylated or M.SssI methylated pGL3-basic vector and the -840/+1 Vsig4 promoter constructs. Error bar, s.e.m. *p < 0.05, **p < 0.01 and NS, p > 0.05 (Student’s t-test). Data are representative of three independent experiments NATURE COMMUNICATIONS 8: 1322 DOI: 10.1038/s41467-017-01327-4 www.nature.com/naturecommunications 9 | | | PBS LPS TNF MALP-2 Poly(I:C) IFN-γ CpG PBS AZAdC AZAdC+LPS AZAdC+TNF AZAdC+MALP-2 AZAdC+Poly(I:C) AZAdC+IFN-γ AZAdC+CpG PBS Poly(I:C) IFN-γ MALP-2 TNF LPS+IFN-γ LPS CpG Untreated Untreated Untreated Untreated LPS TNF MALP-2 Poly(I:C) IFN-γ CpG Count Vsig4 mRNA suppression (suppression) Relative luciferase activity (folds) ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-017-01327-4 improved survival rate (Fig. 8d). These combined data suggest neutrophils is due to reduction in cellular nicotinamide adenine that increasing the expression of VSIG4 might have therapeutic dinucleotide phosphate oxidases activity that results in produc- potentials for fulminant hepatitis and other macrophage- tion of superoxide during the respiratory burst. On the other associated inflammatory disorders (Supplementary Fig. 11). hand, over production of ROS actively participates in the pathogenesis of inflammatory diseases including rheumatoid arthritis, multiple sclerosis, and thyroiditis through activating the inflammatory signaling pathways including mitogen-activated Discussion Macrophage activation relies on metabolic adaptation in response protein kinases (MAPK), NF-κB, and guanylate cyclase . ROS comes from various sources, such as peroxisomes, ubiquinone, to the surrounding micro-environmental stimuli. Macrophage plasticity determines its biological functions in immunity, activities of cytosolic enzymes and uncoupled nitric oxide syn- thases. However, recent data identify that mitochondria are a inflammation, and tissue homeostasis. Defining the mechanisms major source of physiological intracellular ROS that drives regulating macrophage metabolic patterns is critical for under- inflammation . mtROS can be sensed by the NLRP3 inflam- standing the pathology of inflammatory disorders and developing masome, resulting in caspase-1 activation and IL-1β matura- therapeutic interventions . We here demonstrate that VSIG4, a tion . Therefore, high levels of mtROS in vivo would impair B7 family-related protein that is expressed specifically in resting glucose metabolism and insulin sensitivity , and it probably macrophages, is able to inhibit macrophage activation by repro- explains why exaggerated macrophage activation promotes gramming mitochondrial pyruvate oxidation. Vsig4 deficiency pathogenesis of MHV-3-mediated hepatitis . Alternatively, apparently sways macrophage towards activation upon LPS mtROS also activates NF-κB and stabilizes HIF-1α, by thus it exposure in vitro. Conversely, overexpression of Vsig4 suppresses increases the activities of these two transcriptional factors leading M1 gene expression and reduces LPS-induced pyruvate oxidation and mtROS secretion by RAW264.7 cells. Interestingly, Vsig4 to upregulation of macrophage M1 genes . Here, we show that VSIG4 can inhibit pyruvate/acetyl-CoA conversation in deficiency affects the outcomes of inflammatory disorders in −/− animal models. For instance, Vsig4 mice are more susceptible RAW264.7 cells, leading to limitation of oxygen consumption (Fig. 4a–d). Interestingly, decreasing mitochondrial oxidation also to develop HFD-induced obesity and insulin resistance. Fur- −/− leads to the inhibition of LPS-induced mtROS secretion, along thermore, Vsig4 mice exhibit markedly higher mortality over with restriction of Il1b, Il6, and Tnf gene transcription and CD40 MHV-3 viral infection, clearly due to the exacerbated upregulation (Fig. 3). Moreover, PEMs isolated from MHV-3- macrophage-dependent inflammation in vivo. Finally and the −/− −/− infected Vsig4 mice or ATMs collected from HFD-fed Vsig4 most notably, the anti-inflammatory function of VSIG4 has been obese animals all exhibited with more mtROS secretion in vivo as validated by forced overexpression of VSIG4 in vivo, for which we compared to WT littermates (Fig. 4f). Similar to previous stu- show that excessive epigenetic VSIG4 dampens liver tissue dies , inhibition of mtROS activity by DPI successfully prevented inflammation and protects the susceptible mice from MHV-3 virus-induced FH. These studies demonstrate that VSIG4 inhibits LPS-mediated IL-6 production (Fig. 4g). These data suggest that VSIG4 inhibits macrophage activation via posing restriction on macrophage M1-associated inflammatory pathogenesis. It has long been known that ROS plays essential roles in mitochondrial oxidation and mtROS secretion. mtROS derives from pyruvate metabolism during OXPHOS in immune responses of macrophages and neutrophils to pathogens. The bacteria killing capability of activated macrophage and the mitochondria, in which the conversion of cytosolic pyruvate ab Len-Cont. Len-Vsig4 KD #1 #2 #3 #4 #1 #2 #3 #4 2.0 46 PDK2 1.5 S300 43 p-PDH-E1α 1.0 43 PDH 0.5 70 FGL2 0 21 IL-6 17 TNF 17 IL-1β p17 41.6 Actin-β c d Len-Vsig4 100 Len-Cont. Len-Vsig4 (n =8) Len-Cont. (n =8) 0 2 4 6 8 101214161820 Days post infection Fig. 8 Forced overexpression of Vsig4 improves MHV-3-induced hepatitis. C57BL/6 WT mice were infected with lentivirus (10 PFU/mouse) to induce the expression of Vsig4 in vivo, these mice were further infected with MHV-3 at day 6. a Liver Vsig4 gene transcription was analyzed by qRT-PCR at day 6, n = s300 5 per group. b Western blotting for PDK2, p-PDH-E1α , FGL2, and proinflammatory cytokines TNF, IL-6 and IL-1β in liver tissues at 72 h of MHV-3 infection, n = 4 per group. c The architecture of the liver tissues at 72 h of infection was compared by H&E staining, scale bar = 20 μm, n = 5 per group. d The survival was monitored for a total of 20 days. Error bar, s.e.m. a *p < 0.05 was analyzed by Student’s t-test, and d was analyzed by log-rank test. Data are representative of three independent experiments 10 NATURE COMMUNICATIONS 8: 1322 DOI: 10.1038/s41467-017-01327-4 www.nature.com/naturecommunications | | | Len-Cont. Len-Vsig4 mRNA relative (folds) % of survival NATURE COMMUNICATIONS | DOI: 10.1038/s41467-017-01327-4 ARTICLE into mitochondrial acetyl-CoA is partially regulated by the contributes to limiting macrophage M1-associated inflammation, activity of PDH . PDH activity is enhanced via its depho- and its deficiency can cause over inflammatory disorders. sphorylation by phosphopyruvate dehydrogenase phosphatase Identification of mediators that regulate Vsig4 expression will (PDP), and inducing PDP1 expression has been shown to pro- bring insights into understanding the inflammatory response in mote mitochondrial OXPHOS, mtROS production and M1 gene the host. The epigenetic mechanisms, such as DNA methylation expression . Conversely, PDK-mediated phosphorylation of the and histone modification, have recently been shown to mediate 232 293 53 PDH-E1α subunit at 3 serine sites, including Ser , Ser , and gene silencing . DNA methylation is a covalent modifications, in 300 42 Ser , appeared to cause PDH suppression . Recent studies have which DNA methyltransferase catalyze cytosine methylation shown that PDK/PDH axis essentially controls cytokine secretion using S-adenosyl methionine as a methyl donor . It has been in macrophages. For example, overexpression of PDK2 inhibits reported that the expression of tumor suppressor genes in several radiation-induced cytokine expression , whereas macrophages human cancers were inhibited upon DNA methylation at their 37, 55 from Pdk2/4-deficient animals appear to produce less IL-1β and CpG sites within the promoters . We here found that Vsig4 TNF in association with a low level of lactic acid . We here mRNA and protein were decreased dramatically in PEMs that identify that the VSIG4-dependent inhibition of macrophage M1 were treated with inflammatory factors in vitro (Fig. 7c–e). activation relies on PDK2. LPS stimulation resulted in augmen- Conversely, blocking Dnmt3a activity by the inhibitor, 5-aza-2ʹ- tation of mtROS secretion and inflammatory gene expression in deoxycytidine, could effectively prevent proinflammatory factors Pdk2 silenced RAW264.7 cells and Pdk2 deficient BMDMs induced Vsig4 downregulation (Fig. 7e). Further studies demon- (Fig. 5h–j, Supplementary Fig. 8). In agreement, overexpression of strated that the luciferase activity of the M.SssI-methylated Vsig4 Pdk2 results in opposite effects (Fig. 5h–j). Along with previous promoter (-840/+1) was reduced by 53% comparing to the work suggesting that PDK1 is involved in promoting aerobic unmethylated controls, and proinflammatory stimuli were unable glycolysis in macrophages , these combined data highlight a to further inhibit the promoter activity in RAW264.7 cells signaling pathway that governs the fate of macrophages differ- transfected with the M.SssI-methylated Vsig4 promoter con- entiation and function and Pdk status is involved in mediating structs (Fig. 7f). We also found that some CpG sites in Vsig4 metabolic programs for balancing between glycolysis and glucose promoter region (like CpG at -374 bp site) were methylated by oxidation. proinflammatory stimuli as detected by using the Sequenom Substantial evidence demonstrates that obesity is a chronic MassARRAY platform (Supplementary Fig. 10, Supplementary low-grade inflammatory disease . In obesity, adipocytes can Table 1). Together, these data suggest that proinflammatory release proinflammatory mediator like CC chemokine ligand factors repress Vsig4 gene transcription through inducing (CCL)-2, and monocyte chemoattractant molecule (MCP)-1, Dnmt3a activity, causing fast methylation of Vsig4 promoter which induce the recruitment of ATMs . ATM secretes proin- region, for which CpG at -374 bp site, probably with other CpG flammatory cytokines and forms the inflammatory circuit which islands within the Vsig4 promoter region, plays a critical role in blocks the insulin action in adipocytes and leads to insulin regulating Vsig4 transcription. resistance . Here we demonstrate that ATMs isolated from HFD- Recent studies have illustrated that VSIG4 suppresses T-cell −/− fed Vsig4 obese mice have higher levels of proinflammatory cytokine production and causes cell cycle arrest. Injection of factors such as TNF, IFN-γ, and IL-1β (Fig. 1i–k). It was eluci- soluble VSIG4-Ig protein causes a reduction in IFN-γ production dated that proinflammatory cytokines promote obesity-associated by antigen-specific CD8 T cells and down-modulates Th- insulin resistance through activating inflammatory signaling dependent IgG responses in vivo . Additionally, administration pathways, including the stress-responsive c-Jun NH2-terminal of VSIG4-Ig protein prolonged mouse survival in a ConA- kinase (JNK1/2) and AMP-activated protein kinase α2 induced hepatitis model and protected against the pathogenesis of 25, 56 (AMPKα2), inhibitor of κB kinase, and extracellular signal- many inflammatory diseases . One of the explanations for this regulated kinase 1/2 (ERK1/2) as well as MAPKp38, collectively effect is that VSIG4 binds to an as-yet-unidentified inhibitory causing inhibition of serine/threonine phosphorylation of the receptor/ligand on T cells. Alternatively, the surface VSIG4 on 48–50 docking protein IRS-1 . In agreement with previous studies, macrophages and CR1/CR3 on T cells bind to the same multi- −/− we found that Vsig4 obese mice exhibit exacerbated insulin meric C3b or iC3b molecules, thus triggering signals to suppress ser473 57 resistance along with diminished p-IRS-1 and p-Akt in the T-cell activation . Here, we showed that the secretion of cyto- WAT and the liver tissues (Fig. 1h), indicating that Vsig4 defi- kines in T cells that isolated from liver tissues of 72 h post MHV- −/− ciency eliminates the negative control signals, thus allowing 3-infected Vsig4 animals were similar to their WT counter- ATMs to produce proinflammatory cytokines and switching on parts (Supplementary Fig. 3), implicating VSIG4 does not affect HFD-associated insulin tolerance in vivo. T-cell activation in our models. Conversely, we demonstrate that Microorganism infection and tissue injuries trigger the VSIG4 delivers negative feedback signals to macrophages, recruitment of inflammatory macrophages from the circulation resulting in PI3K–Akt–STAT3 dependent PDK2 upregulation, into the affected tissues . The MHV-3 virus provokes a mouse and finally impedes mtROS-dependent M1 macrophage activa- strain-associated severe liver disease that has been used as a tion, suggesting VSIG4 negatively controls macrophage activa- model for investigating human viral fulminant hepatitis. This tion. However, microbeads-C3b did not affect the expression of exacerbation of macrophage-dependent cytokine storm directly PDK2 from LPS-activated VSIG4 THP-1 cells. Additionally, the causes hepatic necrosis and induces lethality in susceptible mouse expression of PDK2 and the secretion of cytokines like IL-6 as 7 −/− strains . In support, normal induction of macrophage apoptosis well as IL1-β were still downregulated in LPS-stimulated C3 can appropriately control local liver tissue fibrinogen deposition BMDMs (Figs. 3h, 6j), these combined data suggest that inhibi- and local tissue injuries in this fulminant hepatitis model .As tion of macrophage activation by VSIG4 is C3b independent. expected, we have found that MHV-3 infection super-induces Further studies are needed to identify the potential interacting −/− production of a panel of inflammatory cytokines from Vsig4 partners for VSIG4. macrophages, and the virus-infected mice exhibit with more In summary, we have demonstrated that VSIG4 down- severe liver necrosis and high mortality (Fig. 2). In contrast, modulates macrophage activation and M1 polarization in forced expression of Vsig4 in susceptible mice provides significant response to inflammatory stimuli in vitro and in vivo. Mechan- protection against MHV-3-induced fulminant hepatitis (Fig. 8). istically, VSIG4 is capable of sending feedback signals in mac- These combined results demonstrate that VSIG4 functionally rophages to activate the PI3K–Akt–STAT3 signaling axis, leading NATURE COMMUNICATIONS 8: 1322 DOI: 10.1038/s41467-017-01327-4 www.nature.com/naturecommunications 11 | | | ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-017-01327-4 to PDK2 upregulation and activation, thus inhibiting mitochon- pyruvate, 2 mM glutamine, pH 7.4) and incubated in a non-CO incubator for 1 h. Four baseline measurements were taken before sequential injection of mitochon- drial pyruvate metabolism, suppressing mtROS secretion and drial inhibitors oligomycin, FCCP, and antimycin (AA) plus rotenone (AR) pro- M1-like gene expression through inducing PDH phosphorylation. vided by the manufacturer (#101706-100, Agilent Technologies). OCR was Therefore, we speculate that enhancing VSIG4 signaling may automatically calculated using the Seahorse XFp software. Every point represents result in beneficial effects on treating inflammatory disorders. an average of three different wells. Methods Immunohistochemistry and immunofluorescence double staining. Paraffin- −/− embedded tissue blocks were cut into 2.5 μm slices and were mounted on Mice. The complement C3 deficient (C3 ) mice (#003641) and the C57BL/6 −/− polylysine-charged glass slides. Endogenous peroxidase activity was blocked by mice were purchased from Jackson Laboratory. The Vsig4 mice were kindly exposure to 3.0% H O for 30 min. Antigen retrieval was performed in a citrate provided by Dr. M. van Lookeren Campagne (Department of Immunology, Gen- 2 2 −/− buffer (pH 6.0) at 120°C for 10 min. Sections were then incubated at 4 °C overnight entech, CA, USA). The Pdk2 mice were provided by Dr. C.R. Harris (Rutgers with anti-mouse FGL2 (#sc-100276, Santa Cruz, 1:100, mouse), anti-Fibrinogen Cancer Institute of New Jersey, USA). All mice were backcrossed ten times onto the (#ab118533, Abcam, Cambridge, England, 1:1000, Rabbit), anti-pro-IL-1β (#12242, B6 background to avoid unpredictable confounders. Specific pathogen-free male Cell Signaling Technology (CST), 1:100, mouse), anti-TNF (#3707, CST, 1:100, and age-matched mice (8–12 weeks old) were used for all experiments. Mice were rabbit), anti-IL-6 (#sc-130326, Santa Cruz, 1:200, mouse), and anti-IFN-γ (#sc- maintained in micro-isolator cages, fed with standard laboratory chow diet and 52557, Santa Cruz, 1:200, rat). After washing, the sections were incubated with the water, and housed in the animal colony at the animal center of the Third Military corresponding secondary antibodies for 2 h at room temperature. The Vecta-stain Medical University (TMMU). All animals received humane care according to the ABC kit (Vector Laboratories, San Diego, CA, USA) was used to perform the criteria outlined in the “Guide for the Care and Use of Laboratory Animals” avidin–biotin complex method according to the manufacturer’s instructions. Sec- prepared by the National Academy of Sciences and published by the National tions incubated with isotype and concentration matched immunoglobulins without Institutes of Health (NIH publication 86–23 revised 1985). All of the in vivo primary antibodies were used as isotype controls. Peroxidase activity was visualized experiments comply with the animal study protocol approved by the ethics com- with the DAB Elite kit (K3465, DAKO), and brown coloration of tissues repre- mittee of TMMU. sented positive staining. S300 To detect p-PDH-E1α and mitochondria co-localization, both BMDMs and Cells. The mouse macrophage cell line RAW264.7, human monocyte cell line PEMs were fixed with 4% paraformaldehyde, permeabilized with 0.1% Saponin in THP-1 and 293T cells were provided by the Cell Institute of the Chinese Academy PBS for 5 min, and blocked with PBS containing 2% BSA for 1 h at 4°C. The cells of Sciences (Shanghai, China). Mouse 17 clone 1 (17CL1) cells were purchased S300 were then stained with rabbit anti-p-PDH-E1α antibodies (#AP1046, 1:100, from ATCC. The cells were cultured in 6-well plates and propagated in DMEM Merk, Temecula, CA, USA) overnight at 4 °C, and then stained with Alexa supplemented with 10% FBS, 100 U/ml penicillin, and 100 μg/ml streptomycin. Fluor488-conjugated donkey anti-rabbit IgG (H+L) highly cross-adsorbed Peritoneal exudative macrophages (PEMs) were harvested and BMDMs induced by secondary antibody (#A-21206, Thermo Scientific, Billerica, MA, USA) for 1 h. M-CSF (#400-28, Peprotech, Rocky Hill, NJ, USA) as described previously . Finally, the sections were incubated with 1 μg/ml DAPI and MitoTracker Red (#M- 7512, Thermo Scientific) at 500 nM for 30 min at 37 °C. The results were analyzed using fluorescence microscopy (Zeiss Axioplan 2). Virus and infection. MHV-3 viruses were expanded in 17CL1 cells to a con- centration of 1 × 10 plaque forming units (PFU)/ml. Mice were received MHV-3 (100 PFU/mouse) via intraperitoneally (i.p.) injection. In some experiments, ELISA and western blotting. The concentration of cytokines, free fatty acid, and C57BL/6 WT mice were infected with lentivirus (1 × 10 PFU/mouse) via intra- triglycerides in the serum or the culture supernatants, the levels of succinate, venous injection to transiently force the expression of Vsig4 in vivo, and these mice pyruvate, acetyl-CoA, triglyceride and PDH enzyme activity in macrophages were were further infected with MHV-3 after 6 days. The virus-infected mice were measured by the ELISA according to the manufacturer’ introductions. ELISA Kits, euthanized on the indicated days. Liver damage was compared by H&E staining including TNF (#EK0527), IL-6 (#EK0411), IL-1β (#EK0394), and IL-12 p40 and cell apoptosis was measured using the Terminal Transferase dUTP Nick End (#EK0932) were from Boster Ltd. (Wuhan, China). FGL2 ELISA Kit was from Uscn Labeling (TUNEL) staining method according to the manufacturer’s instructions Life Science (#SEA512Mu, Wuhan, China). Insulin ELISA Kit was from Millipore (#12156792910, Roche, Mannheim, Germany). The virus titers in liver tissues were (#2617704, Billerica, MA, USA). Pico ProbeAcetyl-CoAFluorometric Assay kit was determined by plaque assay method. Briefly, the supernatant of liver tissue from BioVision (#K317-100, Milpitas, CA, USA). The Free Fatty Acid Quantitation homogenate was ten times step diluted. Mouse 17CL1 cells were seeded in 12-well Kit (#MAK044-1KT), PDH Activity Assay Kit (#MAK183-1KT), Pyruvate Assay plates, when reaches 80% fusion, cells were added with the diluted liver extraction Kit (#MAK071-1KT), Lactate Assay Kit (#MAK064-1KT), Succinate Colorimetric and incubated for 30 min under 37 °C, 5% CO condition. Cells were then added Assay Kit (#MAK184-1KT), and other chemicals were all from Sigma-Aldrich (St. with 1 ml 2% MethyCellulose DMEM medium (100 U/ml penicillin and 100 μg/ml Louis, MO, USA). streptomycin) per well, and then further incubated for 4 days. Cells added with the The expression of GAPDH (#2118, CST, Danvers, MA, USA,1:1000, rabbit), supernatant from uninfected liver tissue homogenate were used as negative con- 307 PDK2 (#sc-14486, Santa Cruz, 1:500; goat), p-IRS-1(p-Ser , #2381, CST, 1:1000, trols, whereas cells treated with the purified MHV-3 virus were used as positive 308 473 rabbit), Akt (#2920, CST, 1:1000, mouse), p-Akt (p-Thy , #4056; p-Ser , #4051, controls. Finally, cells were fixed with 4% paraformaldehyde and the viral titers 293 CST, 1:1000, rabbit), PDH (#3025, CST, 1:1000, rabbit), p-PDH-E1α (#AP1062, were determined by crystal violet staining assay. S300 Merk, 1:2000, rabbit), p-PDH-E1α (#AP1046, Merk, 1:2000, rabbit), STAT3 (#9139, CST, 1:1000, mouse), p-STAT3 (p-Tyr , #9145, CST, 1:2000, rabbit), −/− FGL2 (#sc-100276, Santa Cruz, 1:500, mouse), TNF (#3707, CST, 1:1000, rabbit), Diet intervention. C57BL/6 WT mice and congenic Vsig4 littermates were pro-IL-1β (#12242, CST, 1:1000, mouse), IFN-γ (#sc-52557, Santa Cruz, 1:200, rat), received NCD or HFD (#MD12031, Medicience Ltd., Nanjing, China) starting at and IL-6 (#sc-130326, Santa Cruz, 1:200, mouse) in macrophages or liver tissues as the age of 8 weeks. The body mass was evaluated every week for a total of 10 weeks. well as VAT was measured by western blot. Uncropped western blot images are The distribution of fat tissues in obese mice was made using an in vivo micro- shown in Supplementary Fig. 12. computed tomography scanner (μCT, Quantum FX, Perkin Elmer, Hopkinton, MA, USA). At the end of the feeding experiment, the mice were sacrificed, and blood was collected in EDTA-coated tubes and centrifuged to collect plasma. The Flow cytometry. To measure the mROS superoxide, macrophages were incubated liver and epididymal WAT were dissected, weighed, and immediately frozen in with MitoSOX red (5 μM, Life technologies, Eugene, Oregon, USA) at 1.0 μM for 1 liquid nitrogen. The morphometry of individual fat cells was assessed using digital h in phenol red-free DMEM (Invitrogen). The death cells were excluded firstly by image analysis as described previously . Briefly, microscopic images were digitized staining with LIVE/DEATHFixable Near-IR Ded Cell Stain Kit (Life Technologies). in 24-bit RGB (specimen level pixel size 1.28 × 1.28 μm ). Recognition of fat cells To measure the expression of activation markers on cell surface, suspended cells was initially performed by applying a region growing algorithm on manually were incubated for 1 h at room temperature in dark using fluorescent antibodies indicated seed points, and the minimum Feret diameter was calculated. (anti-B7-H1, anti-B7-DC, anti-B7-H3, and anti-CD40). To detect intracellular proinflammatory cytokines (pro-IL-1β, TNF, and IL-6) expression, macrophages or GTT and ITT experiments. For GTT, the animals were i.p. injected with 2 g/kg T cells were isolated and treated with brefeldin A for 4 h. mAbs were then added, and further incubated for an additional 1 h. all of these fluorescent antibodies were glucose (#G6125, Sigma-Aldrich, St. Louis, MO, USA) after 12 h of fasting, and blood was drawn to measure blood glucose 0, 15, 30, 60, 90 and 120 min after purchased from eBioscience (San Diego, CA, USA). A total of 10,000 live cells were analyzed by FACsAria cytometer (BD, Franklin Lakes, NJ, USA). All the flow injection. For ITT, 0.5 U/kg of insulin (Novolin R, Novo Nordisk) was i.p. injected after 6 h of fasting, and blood was drawn at 0, 15, 30, 45, and 60 min thereafter. cytometry data were analyzed using CellQuest Pro software. Seahorse XFp metabolic flux analysis. The OCR was measured using an XFp Quantitative RT-PCR. Total RNA was extracted from cultured cells or the indi- extracellular analyzer (Agilent Technologies, Santa Clara, CA, USA). Macrophages cated tissues with TRIzol reagent according to the manufacturer’s instructions were seeded at 2.0 × 10 cells/well density in 8-well plates for 5 h to allow adherence (Invitrogen). First-strand cDNA was synthesized with the PrimeScript RT-PCR Kit to the plate. After 2 h of LPS (800 ng/ml) administration, the cells were changed to (Takara, Dalian, China). The expression of mRNA encoding for the indicated genes unbuffered assay media (base medium supplemented with 10 mM glucose, 1 mM was quantified by quantitative (q)RT-PCR with the SYBR Premix ExTaq kit 12 NATURE COMMUNICATIONS 8: 1322 DOI: 10.1038/s41467-017-01327-4 www.nature.com/naturecommunications | | | NATURE COMMUNICATIONS | DOI: 10.1038/s41467-017-01327-4 ARTICLE (Takara) and was normalized to the expression of β-actin. qRT-PCR was per- p < 0.05 was considered a statistically significant difference. All results shown are formed with specific primers (Supplementary Table 2). The results were compared representative of at least three separate experiments. −ΔΔCt by the 2 method. Data availability. The authors declare that all data supporting the findings of this study are available within the article and its Supplementary Information Files or Lentiviral constructs and transduction. The mouseVsig4 (NM_177789) cDNA from the corresponding author upon request. ORF clone (#MR203780) and Pdk2 (NM_133667) cDNA ORF clone (#MG206400) were purchased from OriGene Technologies, Inc. (Rockville, MD, USA). The whole gene expression cDNA for Vsig4, Pdk2, the truncation mutants of Vsig4, and Vsig4 Received: 7 September 2016 Accepted: 11 September 2017 site-directed mutagenesis were further amplified with specific primers (Supple- mentary Table 3). cDNA was cloned into the pCDH-MCS-T2A-copGFP-MSCV (CD523A-1) vector. This vector was mutated using the QuickChange site-directed mutagenesis kit II (Stratagene, Santa Clara, CA, USA). The lentiviral packaging vectors-psPAX2 and pVSVG were purchased from Addgene (Cambridge, MA, USA). The psPAX2 plasmids (2 μg), the expression vectors (2 μg) and the pVSVG References plasmids (2 μg) were cotransfected into 293T cells, and the virus supernatants were 1. Wynn, T. A. & Vannella, K. M. Macrophages in tissue repair, regeneration, and collected after 48 h (2,000 rpm/min, 3 min). RAW264.7 cells were transduced with unconcentrated virus supernatant overnight in the presence of 8 mg/ml polybrene fibrosis. Immunity 44, 450–462 (2016). and selected in puromycin (0.5 mg/ml). The expression of VSIG4 and PDK2 was 2. Murray, P. J. & Wynn, T. A. Protective and pathogenic functions of measured by western blot. macrophage subsets. Nat. Rev. Immunol. 11, 723–737 (2011). 3. Sica, A. & Mantovani, A. Macrophage plasticity and polarization: in vivo veritas. J. Clin. Investig. 122, 787–795 (2012). Pdk2 and Stat3 shRNA silencing. Lentiviral constructs with shRNAs directed 4. Murray, P. J. et al. Macrophage activation and polarization: nomenclature and against mouse Pdk2 and Stat3 in the pGV112 vectors, which were prepared by experimental guidelines. Immunity 41,14–20 (2014). GeneChem (Shanghai, China). Primers are presented in Supplementary Table 4. 5. Olefsky, J. M. & Glass, C. K. Macrophages, inflammation, and insulin Lentivirus was prepared by transient transfection of 293T cells with transfer vectors resistance. Annu. Rev. Physiol. 72, 219–246 (2010). along with third-generation packaging constructs (pHelper 1.0 and pHelper 2.0). 6. Weisberg, S. P. et al. Obesity is associated with macrophage accumulation in The viral titers were determined with serial dilution of virus-containing media on adipose tissue. J. Clin. Investig. 112, 1796–1808 (2003). NIH3T3 cells. RAW264.7 cells were transfected with unconcentrated virus 7. Pope, M. et al. Pattern of disease after murine hepatitis virus strain 3 infection supernatant overnight in the presence of 8 mg/ml polybrene and selected in pur- correlates with macrophage activation and not viral replication. J. Virol. 69, omycin (0.5 mg/ml). 5252–5260 (1995). 8. Zhang, L. & Wang, C. C. Inflammatory response of macrophages in infection. Hepatobiliary Pancreat. Dis. Int. 13, 138–152 (2014). Chromatin immunoprecipitation and qPCR. BMDMs underwent cross-linking for 10 min with 1% formaldehyde in medium. Chromatin fragments were pre- 9. Ganeshan, K. & Chawla, A. Metabolic regulation of immune responses. Annu. pared, followed by immunoprecipitation with anti-Dnmt3a (#ab2850, Abcam, Rev. Immunol. 32, 609–634 (2014). 1:1000, Rabbit), anti-STAT3 (#ab119352, Abcam, 1:1000, mouse) or rabbit/mosue 10. Jha, A. K. et al. Network integration of parallel metabolic and transcriptional IgG isotype mAbs and coupled to Dynabeads Protein G. DNA was phenol/ data reveals metabolic modules that regulate macrophage polarization. chloroform extracted and ethanol precipitated in the presence of glycogen. DNA Immunity 42, 419–430 (2015). was purified with a PCR purification kit (#28104, Qiagen). qPCR was performed 11. Yeung, F. et al. Modulation of NF-kappaB-dependent transcription and cell with special primers (Supplementary Table 5) flanking the putative Dnmt3a- and survival by the SIRT1 deacetylase. EMBO J. 23, 2369–2380 (2004). STAT3-binding sites. The input DNA was an aliquot of sheared chromatin before 12. Kauppinen, A., Suuronen, T., Ojala, J., Kaarniranta, K. & Salminen, A. immunoprecipitation and was used for normalization of the samples to the amount Antagonistic crosstalk between NF-kappaB and SIRT1 in the regulation of chromatin added to each ChIP. of inflammation and metabolic disorders. Cell Signal. 25, 1939–1948 (2013). 13. Tannahill, G. M. et al. Succinate is an inflammatory signal that induces IL-1beta Vsig4 gene promoter methylation analyses. The Sequenom MassARRAY plat- through HIF-1alpha. Nature 496, 238–242 (2013). form (CapitalBio, Beijing, China) was used to perform the quantitative methylation 14. West, A. P. et al. TLR signalling augments macrophage bactericidal activity analysis of Vsig4. This system uses matrix-assisted laser desorption/ionization through mitochondrial ROS. Nature 472, 476–480 (2011). time-of-flight (MALDI-TOF) mass spectrometry in combination with RNA base- 15. Zhou, R., Yazdi, A. S., Menu, P. & Tschopp, J. A role for mitochondria in specific cleavage (MassCLEAVE). A detectable pattern is then analyzed for its NLRP3 inflammasome activation. Nature 469, 221–225 (2011). methylation status. PCR primers were designed with Methprimer (http:// 16. Wen, H. et al. Fatty acid-induced NLRP3-ASC inflammasome activation epidesigner.com) and were showed in (Supplementary Table 6). The spectra interferes with insulin signaling. Nat. Immunol. 12, 408–415 (2011). methylation ratios were generated by Epityper software version 1.0 (Sequenom, San 17. Michiels, C., Minet, E., Mottet, D. & Raes, M. Regulation of gene expression by Diego, CA). oxygen: NF-kappaB and HIF-1, two extremes. Free Radic. Biol. Med. 33, 1231–1242 (2002). Vsig4 gene promoter luciferase reporter assay. A840bp (-840/+1) fragment of 18. Papandreou, I., Cairns, R. A., Fontana, L., Lim, A. L. & Denko, N. C. HIF-1 Vsig4 gene promoter was cloned into the pGL3-basic expression vector using mediates adaptation to hypoxia by actively downregulating mitochondrial primers described in Supplementary Table 7. The pGL3-basic and the -840/+1 oxygen consumption. Cell Metab. 3, 187–197 (2006). Vsig4 promoter constructs were fully methylated in vitro by M.SssI following the 19. Xu, J. et al. NOTCH reprograms mitochondrial metabolism for manufacturer’s instructions (New England Biolabs (NEB), Beijing, China). The proinflammatory macrophage activation. J. Clin. Investig. 125, 1579–1590 completeness of methylation was checked by measuring the extent of protection (2015). from digestion by the restriction enzymes BstUI (NEB). RAW264.7 cells were 20. Vogt, L. et al. VSIG4, a B7 family-related protein, is a negative regulator of T cultured in 12-well plates and were transfected with these plasmids by using cell activation. J. Clin. Investig. 116, 2817–2826 (2006). amaxanucleofector (4D-Nucleofector, Lonza, Allendale, NJ, USA). At 48 h after 21. Helmy, K. Y. et al. CRIg: a macrophage complement receptor required for transfection, cells were further treated with proinflammatory mediators for an phagocytosis of circulating pathogens. Cell 124, 915–927 (2006). additional 12 h. Thereafter, cells were washed with PBS and lysed in Reporter Lysis 22. Fu, W., Wojtkiewicz, G., Weissleder, R., Benoist, C. & Mathis, D. Early Buffer (Invitrogen). Luciferase reporter activities were measured in triplicate using window of diabetes determinism in NOD mice, dependent on the complement the Dual-Luciferase reporter assay system (Promega, Madison, WI, USA) receptor CRIg, identified by noninvasive imaging. Nat. Immunol. 13, 361–368 according to the manufacturer’s protocol, and quantified using the GloMax 96-well (2012). plate luminometer (Promega). The firefly luciferase to Renilla luciferase ratios were 23. Katschke, K. J. Jr et al. A novel inhibitor of the alternative pathway of determined and were defined as the relative luciferase activity. Results are shown as complement reverses inflammation and bone destruction in experimental the mean ± s.e.m. of a representative experiment performed in triplicate. To arthritis. J. Exp. Med. 204, 1319–1325 (2007). examine the transfection efficiency, RAW264.7 cells were transfected with a 24. Chen, M., Muckersie, E., Luo, C., Forrester, J. V. & Xu, H. Inhibition of the pmaxGFP control vector and transfection efficiency was assessed by in situ GFP alternative pathway of complement activation reduces inflammation in expression according to manufacturer’s protocol (Lonza). experimental autoimmune uveoretinitis. Eur. J. Immunol. 40, 2870–2881 (2010). 25. Jung, K. et al. Protective role of V-set and immunoglobulin domain-containing Statistical analysis. Survival data from in vivo experiments were analyzed by a log-rank test performed on curves generated by GraphPad Prism 4.03 Software 4 expressed on kupffer cells during immune-mediated liver injury by inducing tolerance of liver T- and natural killer T-cells. Hepatology 56, 1838–1848 (MacKiev). For all other analysis, two-tailed, unpaired Student’s t-tests with a 95% confidence interval performed on graphs generated in GraphPad Prism were used. (2012). NATURE COMMUNICATIONS 8: 1322 DOI: 10.1038/s41467-017-01327-4 www.nature.com/naturecommunications 13 | | | ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-017-01327-4 26. Stienstra, R. et al. Inflammasome is a central player in the induction of obesity 50. Li, P. P. et al. LTB4 promotes insulin resistance in obese mice by acting on and insulin resistance. Proc. Natl. Acad. Sci. USA 108, 15324–15329 (2011). macrophages, hepatocytes and myocytes. Nat. Med. 21, 239–247 (2015). 27. Uysal, K. T., Wiesbrock, S. M., Marino, M. W. & Hotamisligil, G. S. Protection 51. Chow, S. H., Deo, P. & Naderer, T. Macrophage cell death in microbial from obesity-induced insulin resistance in mice lacking TNF-alpha function. infections. Cell. Microbiol. 18, 466–474 (2016). Nature 389, 610–614 (1997). 52. Yang, C. et al. Expression of B and T lymphocyte attenuator (BTLA) in 28. Marsden, P. A. et al. The Fgl2/fibroleukin prothrombinase contributes to macrophages contributes to the fulminant hepatitis caused by murine hepatitis immunologically mediated thrombosis in experimental and human viral virus strain-3. Gut 62, 1204–1213 (2013). hepatitis. J. Clin. Investig. 112,58–66 (2003). 53. Baylin, S. B., Belinsky, S. A. & Herman, J. G. Aberrant methylation of gene 29. Guo, S. et al. TheNLRP3 Inflammasome and IL-1beta accelerate promoters in cancer---concepts, misconcepts, and promise. J. Natl Cancer. Inst. immunologically mediated pathology in experimental viral fulminant hepatitis. 92, 1460–1461 (2000). PLoS Pathog. 11, e1005155 (2015). 54. Jaenisch, R. & Bird, A. Epigenetic regulation of gene expression: how the 30. Liu, J. et al. C5aR, TNF-alpha, and FGL2 contribute to coagulation and genome integrates intrinsic and environmental signals. Nat. Genet. 33, 245–254 complement activation in virus-induced fulminant hepatitis. J. Hepatol. 62, (2003). 354–362 (2015). 55. Prokhortchouk, E. & Defossez, P. A. The cell biology of DNA methylation in 31. Zhu, L., Zhao, Q., Yang, T., Ding, W. & Zhao, Y. Cellular metabolism and mammals. Biochim. Biophys. Acta 1783, 2167–2173 (2008). macrophage functional polarization. Int. Rev. Immunol. 34,82–100 (2015). 56. He, J. Q., Wiesmann, C. & van Lookeren Campagne, M. A role of macrophage 32. Balaban, R. S., Nemoto, S. & Finkel, T. Mitochondria, oxidants, and aging. Cell complement receptor CRIg in immune clearance and inflammation. Mol. 120, 483–495 (2005). Immunol. 45, 4041–4047 (2008). 33. Patel, M. S. & Korotchkina, L. G. Regulation of mammalian pyruvate 57. Zang, X. & Allison, J. P. To be or not to be B7. J. Clin. Investig. 116, 2590–2593 dehydrogenase complex by phosphorylation: complexity of multiple (2006). phosphorylation sites and kinases. Exp. Mol. Med. 33, 191–197 (2001). 34. Singh, S. S. et al. Targeting the PI3K/Akt signaling pathway in gastric Acknowledgements carcinoma: a reality for personalized medicine? World J. Gastroenterol. 21, Dr. Bin Li (Institute Pasteur of Shanghai, Chinese Academy of Sciences) gave invaluable 12261–12273 (2015). suggestions. This work was supported by the General Program of National Natural 35. Woolfrey, K. M. & Dell’Acqua, M. L. Coordination of protein phosphorylation Science Foundation of China (NSFC, No. 81361120400, 81222023 and 81771691), and and dephosphorylation in synaptic plasticity. J. Biol. Chem. 290, 28604–28612 The National Key Research and Development Program of China (2016YFA0502204). (2015). 36. Durant, L. et al. Diverse targets of the transcription factor STAT3 contribute to T cell pathogenicity and homeostasis. Immunity 32, 605–615 (2010). Author contribution 37. Nagase, H. & Ghosh, S. Epigenetics: differential DNA methylation in Y.C. and Y.W. designed the research and analyzed the data. J.L., S.G., and X.H. per- mammalian somatic tissues. FEBS J. 275, 1617–1623 (2008). formed most experiments. Q.N. and W.Y. prepared MHV-3 virus. The manuscript was 38. Kamata, H. et al. Reactive oxygen species promote TNFalpha-induced death prepared by L.Z., and Y.C. and Z.F. conducted the plasmid mutation. B.D. and C.Y. and sustained JNK activation by inhibiting MAP kinase phosphatases. Cell 120, conducted HFD-induced obesity. 649–661 (2005). 39. Sorbara, M. T. & Girardin, S. E. Mitochondrial ROS fuel the inflammasome. Additional information Cell Res. 21, 558–560 (2011). Supplementary Information accompanies this paper at doi:10.1038/s41467-017-01327-4. 40. Bulua, A. C. et al. Mitochondrial reactive oxygen species promote production of proinflammatory cytokines and are elevated in TNFR1-associated periodic Competing interests: The authors declare no competing financial interests. syndrome (TRAPS). J. Exp. Med. 208, 519–533 (2011). 41. Jha, M. K., Lee, I. K. & Suk, K. Metabolic reprogramming by the pyruvate Reprints and permission information is available online at http://npg.nature.com/ dehydrogenase kinase-lactic acid axis: Linking metabolism and diverse reprintsandpermissions/ neuropathophysiologies. Neurosci. Biobehav. Rev. 68,1–19 (2016). 42. Peterson, R. T. & Schreiber, S. L. Kinase phosphorylation: keeping it all in the Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in family. Curr. Biol. 9, R521–R524 (1999). published maps and institutional affiliations. 43. Yoo, H. et al. Pyruvate metabolism: a therapeutic opportunity in radiation- induced skin injury. Biochem. Biophys. Res. Commun. 460, 504–510 (2015). 44. Jha, M. K. et al. Metabolic connection of inflammatory pain: pivotal role of a pyruvate dehydrogenase kinase-pyruvate dehydrogenase-lactic acid axis. J. Open Access This article is licensed under a Creative Commons Neurosci. 35, 14353–14369 (2015). Attribution 4.0 International License, which permits use, sharing, 45. Tan, Z. et al. Pyruvate dehydrogenase kinase 1 participates in macrophage adaptation, distribution and reproduction in any medium or format, as long as you give polarization via regulating glucose metabolism. J. Immunol. 194, 6082–6089 appropriate credit to the original author(s) and the source, provide a link to the Creative (2015). Commons license, and indicate if changes were made. The images or other third party 46. Vandanmagsar, B. et al. The NLRP3 inflammasome instigates obesity-induced material in this article are included in the article’s Creative Commons license, unless inflammation and insulin resistance. Nat. Med. 17, 179–188 (2011). indicated otherwise in a credit line to the material. If material is not included in the 47. Kanda, H. et al. MCP-1 contributes to macrophage infiltration into adipose article’s Creative Commons license and your intended use is not permitted by statutory tissue, insulin resistance, and hepatic steatosis in obesity. J. Clin. Investig. 116, regulation or exceeds the permitted use, you will need to obtain permission directly from 1494–1505 (2006). 48. Traves, P. G. et al. Relevance of the MEK/ERK signaling pathway in the the copyright holder. To view a copy of this license, visit http://creativecommons.org/ metabolism of activated macrophages: a metabolomic approach. J. Immunol. licenses/by/4.0/. 188, 1402–1410 (2012). 49. Wu, Y. et al. Activation of AMPK alpha 2 in adipocytes is essential for nicotine- © The Author(s) 2017 induced insulin resistance in vivo. Nat. Med. 21, 373–382 (2015). 14 NATURE COMMUNICATIONS 8: 1322 DOI: 10.1038/s41467-017-01327-4 www.nature.com/naturecommunications | | | http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Nature Communications Springer Journals

VSIG4 inhibits proinflammatory macrophage activation by reprogramming mitochondrial pyruvate metabolism

Loading next page...
 
/lp/springer-journals/vsig4-inhibits-proinflammatory-macrophage-activation-by-reprogramming-Ul43WN6YK9

References (59)

Publisher
Springer Journals
Copyright
Copyright © 2017 by The Author(s)
Subject
Science, Humanities and Social Sciences, multidisciplinary; Science, Humanities and Social Sciences, multidisciplinary; Science, multidisciplinary
eISSN
2041-1723
DOI
10.1038/s41467-017-01327-4
pmid
29109438
Publisher site
See Article on Publisher Site

Abstract

ARTICLE DOI: 10.1038/s41467-017-01327-4 OPEN VSIG4 inhibits proinflammatory macrophage activation by reprogramming mitochondrial pyruvate metabolism 1 1 1 1 1 1 2 2 Jialin Li , Bo Diao , Sheng Guo , Xiaoyong Huang , Chengying Yang , Zeqing Feng , Weiming Yan , Qin Ning , 3 1 1 Lixin Zheng , Yongwen Chen & Yuzhang Wu Exacerbation of macrophage-mediated inflammation contributes to pathogenesis of various inflammatory diseases, but the immunometabolic programs underlying regulation of mac- rophage activation are unclear. Here we show that V-set immunoglobulin-domain-containing 4 (VSIG4), a B7 family-related protein that is expressed by resting macrophages, inhibits −/− macrophage activation in response to lipopolysaccharide. Vsig4 mice are susceptible to high-fat diet-caused obesity and murine hepatitis virus strain-3 (MHV-3)-induced fulminant hepatitis due to excessive macrophage-dependent inflammation. VSIG4 activates the PI3K/ Akt–STAT3 pathway, leading to pyruvate dehydrogenase kinase-2 (PDK2) upregulation and subsequent phosphorylation of pyruvate dehydrogenase, which results in reduction in pyr- uvate/acetyl-CoA conversion, mitochondrial reactive oxygen species secretion, and macro- phage inhibition. Conversely, interruption of Vsig4 or Pdk2 promotes inflammation. Forced expression of Vsig4 in mice ameliorates MHV-3-induced viral fulminant hepatitis. These data show that VSIG4 negatively regulates macrophage activation by reprogramming mitochon- drial pyruvate metabolism. 1 2 Institute of Immunology, PLA, Third Military Medical University, Chongqing, 400038, China. Institute of Infectious Disease, Tongji Hospital of Tongji Medical College, Huazhong University of Science and Technology, Wuhan, 430030, China. Laboratory of Immunology, National Institute of Allergy and Infectious Diseases, NIH, Bethesda, Maryland, MD 20892, USA. Jialin Li and Bo Diao contributed equally to this work. Correspondence and requests for materials should be addressed to Y.C. (email: yongwench@163.com) or to Y.W. (email: wuyuzhang@tmmu.edu.cn) NATURE COMMUNICATIONS 8: 1322 DOI: 10.1038/s41467-017-01327-4 www.nature.com/naturecommunications 1 | | | ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-017-01327-4 acrophages are essential for innate immunity owing to that VSIG4 can deliver anti-inflammatory signals. Here, we show functions in host defense, tissue development and that VSIG4 antagonizes activation signals in macrophages by Mhomeostasis . Macrophage functional disparity is stimulating PI3K/Akt–STAT3 cascades, augmenting expression attributed to two distinct subgroups, namely M1 (classically of pyruvate dehydrogenase kinase-2 (PDK2), and inhibiting activated) and M2 (alternatively activated) macrophages .M1 pyruvate dehydrogenase (PDH) activity via phosphorylation. macrophages have been implicated in initiating and sustaining Therefore, VSIG4 restricts pyruvate metabolism in the mito- inflammation in response to INF-γ and/or lipopolysaccharide chondria during oxidative phosphorylation (OXPHOS), resulting (LPS), whereas IL-4 or IL-13 polarized M2 macrophages seem to in suppression of mtROS secretion and M1 differentiation. have immunoregulatory functions in parasitic infections, tissue Conversely, Vsig4 or Pdk2 deficiency enhances macrophage −/− inflammation, remodeling of damaged tissue, and tumor pro- activation. Vsig4 mice are more susceptible to HFD-induced gression . Uncontrolled M1 activation can cause tissue damage obesity in association with insulin resistance. These mice manifest and pathogenesis in inflammatory diseases, including athero- with more severe liver damage and mortality as a result of MHV- sclerosis, obesity, diabetes, rheumatoid arthritis and hepatitis .In 3 infection, probably owing to overloading of macrophage- a model of mouse obesity caused by high-fat diet (HFD), the mediated inflammation in vivo. Interestingly, forced over- accumulation of adipose tissue macrophages (ATMs) within the expression of Vsig4 ameliorates MHV-3-induced viral fulminant white adipose tissue (WAT) involves remodeling of the enlarged hepatitis. These data identify an inhibitory function of VSIG4 in WAT and induction of insulin resistance via secreting proin- macrophage-mediated inflammation. 5, 6 flammatory mediators IL-1β and TNF . Moreover, in a viral fulminant hepatitis model, murine hepatitis virus strain-3 (MHV- 3) infection induces a macrophage-dependent cytokine storm of Results −/− IL-1, TNF, TGF-β, leukotriene B4, and pro-coagulant fibrinogen- Vsig4 mice are more susceptible to HFD-induced obesity. like protein-2 (FGL2), which causes fibrin deposition in the liver Macrophages actively contribute to the pathogenesis of diet- 5, 6 and results in acute hepatic necrosis and lethality of susceptible induced obesity , making them seemingly a good model for mice . Therefore, attenuation of macrophage-mediated inflam- examination of the biological functions of VSIG4. For this, age- −/− mation is a plausible strategy for treating inflammatory disorders. matched C57BL/6 wild type (WT) and Vsig4 mice were fed a −/− The regulation of macrophage activation has been extensively HFD. Interestingly, we found that Vsig4 mice gained sig- studied, with evidence suggesting the involvement of multiple nificantly more body masses than their WT controls after 5 weeks intracellular signaling regulators, including membrane molecules, of HFD feeding (Fig. 1a). However, no significant differences of small noncoding RNAs, microRNAs and epigenetic-associated body masses between the two groups fed a normal chow diet mechanisms . Evidence from metabolic screening and microarray (NCD) (Supplementary Fig. 1a). Analytical microCT (μCT) analyses shows that LPS-activated macrophages have alterations imaging revealed an increase in fat mass throughout the bodies of −/− in mitochondrial metabolites, indicating that reprogramming of the Vsig4 obese mice, with substantially more abundant −/− mitochondrial metabolism may be involved in the regulation of visceral fat (Fig. 1b). Vsig4 obese mice also manifested with 9, 10 + macrophage activation . For example, NAD , as a mitochon- significant increases in abdominal wall fat and perirenal fat drial intermediate metabolite, can inhibit inflammation through compared to the WT controls. In parallel, we also observed a 11, 12 inactivating transcription factor NF-κB . Conversely, succi- substantial increase in the serum levels of triglyceride, cholesterol nate, another important mitochondrial metabolite, accumulates and free fatty acid, which correlated to increases in body fat in −/− −/− in LPS-activated macrophages and promotes the transcription of Vsig4 obese mice (Fig. 1c). Furthermore, Vsig4 obese mice Il1b by stabilizing hypoxia-inducible factor 1α (HIF-1α) . tended to develop with high liver triglyceride levels and steatosis Moreover, mitochondrial reactive oxygen species (mtROS) can (Fig. 1d), as well as enlarge adipocytes (Fig. 1e). These data imply activate the NLRP3 inflammasome and trigger bioprocessing of that Vsig4 deficiency renders mice susceptible to HFD-induced proinflammatory cytokines including pro-IL-1β and pro-IL-18 in obesity. 14, 15 5 macrophages . Therefore, inhibition of autophagy, which Obesity is often associated with insulin resistance , a fact that −/− impairs the removal of ROS-generating mitochondria and causes led us to postulate that Vsig4 mice might have a disturbed NLRP3 inflammasome activation and IL-1β secretion, attenuates glucose metabolism. We next examined the blood glucose levels 16 −/− glucose tolerance and insulin sensitivity . Additionally, mtROS of these obese mice. Vsig4 obese mice exhibited statistically can induce M1 activation through activating NF-κB and stabi- significant elevations in blood glucose relative to WT littermates 17, 18 lizing HIF-1α . Therefore, suppression of mtROS secretion after 15 h of fasting. Moreover, the 5-h-fasting insulin levels were can mitigate pathogenesis in alcoholic steatohepatitis and reduce also substantially elevated (Fig. 1f). An oral glucose tolerance test 19 −/− lethality in endotoxin-mediated fulminant hepatitis . Never- (GTT) illustrated that Vsig4 obese mice had significantly theless, the mechanisms underlying reprogramming of mito- higher serum levels of glucose and insulin in response to the chondrial metabolism during macrophage activation are unclear. glucose load compared to their WT littermate controls, suggesting V-set immunoglobulin-domain-containing 4 (VSIG4) is a a severely impaired glucose metabolism (Fig. 1g). Nevertheless, membrane protein belonging to complement receptor of the the blood glucose metabolism was similar between the two groups 20, 21 immunoglobulin superfamily (CRIg) . By binding comple- under NCD conditions (Supplementary Fig. 1b, c). The insulin ment component C3b, VSIG4 mediates clearance of C3b- tolerance test (ITT) also indicated the existence of a significantly −/− opsonized pathogens, such as Listeria monocytogenes and Sta- more resistance in Vsig4 obese mice compared to the WT phylococcus aureus . The expression of VSIG4 is restricted to controls (Fig. 1g). Additionally, as compared to WT controls, −/− tissue macrophages, including peritoneal macrophages and liver- western blot analyses revealed that Vsig4 obese mice had residential Kupffer cells. Moreover, VSIG4 marks a subset of diminished phosphorylation of IRS-1 (p-IRS-1) and phosphor- 22 ser473 macrophages that associates with diabetes resistance . VSIG4 ylation of Akt (p-Akt ) in the visceral adipose tissue (VAT), can functionally inhibit IL-2 production and T-cell proliferation muscle and liver tissues after administrated with insulin (Fig. 1h). by binding an unidentified T-cell ligand or receptor . Interest- These data indicate that Vsig4 deficiency promotes obesity- ingly, a VSIG4-Fc fusion protein seems to protect against devel- associated insulin resistance. opment of experimental arthritis , experimental autoimmune To better understand the mechanisms by which the VSIG4 24 25 uveoretinitis , and immune-mediated liver injuries , suggesting pathway prevents weight gain during HFD consumption, we 2 NATURE COMMUNICATIONS 8: 1322 DOI: 10.1038/s41467-017-01327-4 www.nature.com/naturecommunications | | | NATURE COMMUNICATIONS | DOI: 10.1038/s41467-017-01327-4 ARTICLE −/− addressed the food intake of these mice. There were no obvious Vsig4 mice exacerbate MHV-3-induced fulminant hepatitis. changes in food intake and stool output between the WT and We further investigated the potential role of Vsig4 in viral ful- −/− Vsig4 mice under HFD conditions (Supplementary Fig. 2), minant hepatitis caused by MHV-3 infection, in which the virus- implying that the differences in weight gain are due to reasons induced exaggerated inflammation causes severe pathogenesis other than food consumption. We therefore examined the status largely due to the M1 macrophage-dependent “cytokine storm” . −/− of the ATMs. Compared to WT littermates, flow cytometric assay Vsig4 and congenic C57BL/6 WT littermates were infected −/− −/− showed that the ATMs from Vsig4 obese mice expressed with MHV-3 (100 PFU/mouse). Noticeably, Vsig4 mice died higher amounts of proinflammatory factors (like pro-IL-1β, IFN- rapidly following MHV-3 infection in contrast to WT littermates γ, and TNF), which have been described to actively participate (log-rank test, p = 0.0339, Fig. 2a). H&E staining showed that the −/− into the pathogenesis of HFD-caused obesity and insulin infected Vsig4 animals had more severe liver necrosis and 26, 27 resistance (Fig. 1i). In parallel, dramatically higher levels of hepatocyte apoptosis at 48 h and 72 h of infection (Fig. 2b), along −/− these cytokines were accumulated in the VAT of Vsig4 mice, with significantly higher levels of serum alanine aminotransferase as detected by qRT-PCR (Fig. 1j) and western blot (Fig. 1k), (ALT) and aspartate aminotransferase (AST), the two liver respectively. These data suggest that Vsig4 deficiency initiates damage indicating enzymes released into the blood compared macrophage-mediated inflammation, which triggers HFD- with WT controls (Fig. 2c). Plaque assay data also showed that induced obesity and insulin resistance. Vsig4 deficiency promoted virus replication in liver tissues a bc –/– WT Vsig4 35 WT (n =5) * ** * –/– ** 5 * 2.5 *** 1.25 * 5 5 Vsig4 (n =5) **** ** ** 2.0 1.00 4 4 1.5 3 3 3 0.75 2 2 2 1.0 0.50 1 1 1 0.5 0.25 15 0 0 0 0 –/– WT Vsig4 Weeks of HFD feeding de –/– –/– WT Vsig4 WT Vsig4 5 *** * * 4 20 3 15 2 10 1 5 0 0 0 f h –/– –/– Vsig4 –/– Vsig4 WT 25 Vsig4 12 10 ** 6 –/– *** Vsig4 12 KD –+ + +–+ + + Insulin 5 8 * * 180 p-IRS-1 6 9 WT 6 473 3 60 pAKT 10 4 WT 2 WT 60 AKT 5 2 3 GTT GTT ITT 37 GAPDH 0 0 0 0 0 15306090120 0 15 30 45 60 0 15 30 45 60 180 p-IRS-1 Times (min) Times (min) Times (min) p-AKT 60 AKT i k GAPDH Il-1 Tnf IL-6 –/– WT Vsig4 p-IRS-1 p = 0.14 180 5 * 2.0 * 3 KD #1 #2 #3 #1 #2 #3 60 p-AKT 1.5 17 TNF AKT 3 60 25 IFN-γ 1.0 37 GAPDH 17 IL1-β 0.5 GAPDH Pro-IL-1β IFN-γ TNF 0 0 –/– WT Vsig4 –/– WT Vsig4 −/− −/− Fig. 1 Vsig4 mice are more susceptible to HFD-induced obesity with insulin resistance. Eight-week-old male Vsig4 mice and age-matched C57BL/6 WT controls were fed a HFD. a Body weight was measured and compared. The obese mice were sacrificed after 10 weeks of HFD feeding. b Fat distribution was detected by μCT. Yellow indicates subcutaneous fat and brown indicates that visceral fat. c Measurement of abdominal wall fat, perirenal fat, serum triglyceride, cholesterol, and free fatty acid. d Representative liver H&E staining (left), and intrahepatic triglyceride contents (right), scale bar = 20 μm, n = 10 per group. e Representative the architecture of adipose tissues stained by H&E (left), adipocyte size and cell numbers was calculated (right), scale bar = 20 μm, n = 10 per group. f The 15-h-fasting blood glucose levels and 5-h-fasting serum insulin levels. g GTT and ITT were performed in theses obese mice, ser473 n = 6 per group. h Western blot of the AKT, p-Akt , and p-IRS-1 in VAT, muscle and liver tissues of obese mice after 4 min of insulin administration, n = 4 per group. ATMs were isolated from obese mice. i Flow cytometry analyzing pro-IL-1β, IFN-γ, and TNF. Cytokines in VAT were detected by j qRT-PCR and k western blot. Error bar, s.e.m. *p < 0.05, **p < 0.01 and ***p < 0.001 (Student’s t-test). Data are representative of five (a) and three (c–i) independent experiments NATURE COMMUNICATIONS 8: 1322 DOI: 10.1038/s41467-017-01327-4 www.nature.com/naturecommunications 3 | | | WT –/– Vsig4 WT –/– Vsig4 WT –/– Vsig4 WT –/– Vsig4 WT –/– Vsig4 Count 15 h-fasting glucose (mM/l) Body mass (g) 5 h-fasting insulin (ng/ml) 7 mRNA relative (folds) Blood glucose (mM/l) Triglyceride (U/mL) mRNA relative (folds) Insulin (ng/ml) mRNA relative (folds) Abdominal wall fat (g) Blood glucose (mM/l) Perrienal fat (g) Cholesterol (U/ml) Adipocyte size (μm ) Liver Muscle VAT Triglyceride (U/ml) 5 2 Cell numb. (x10 )/cm Free fatty acid (U/ml) ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-017-01327-4 (Fig. 2d). This suggests that VSIG4 is capable of attenuating (Fig. 2e), western blot (Fig. 2g), and immunohistochemistry MHV-3-induced pathogenesis. (Supplementary Fig. 4). Therefore, dramatically higher levels of The macrophage-derived FGL2 and proinflammatory cyto- these proinflammatory cytokines were accumulated in the virus- −/− −/− kines including TNF, IL-1 and IL-6, play essential roles in the infected Vsig4 serum (Fig. 2h). Finally, Vsig4 mice 28–30 pathogenesis of MHV-3-induced fulminant hepatitis .We responded with severe fibrinogen formation, leading to increased therefore examined the expression of these mediators in liver coagulation and necrosis post infection (Supplementary peritoneal exudate macrophages (PEMs) and liver tissues. Fig. 4). VSIG4 has been described to be a negative regulator of T- Although Vsig4 deficiency did not seem to affect the basal cell activation , nevertheless, the secretion of proinflammatory expression level of these factors before viral infection (Supple- cytokines (TNF and IFN-γ), and the expression of activated + + mentary Fig. 3), qRT-PCR data showed that MHV-3 infection markers (CD25 and CD69) from CD4 as well as CD8 T cells in −/− −/− super-induced Vsig4 PEMs to express Fgl2, Tnf, Il-1β, and Il-6 72 h of MHV-3-infected Vsig4 livers was similar to their WT (Fig. 2e). Flow cytometric data also confirmed at protein levels controls (Supplementary Fig. 5). These results clearly demon- that inflammatory cytokines like pro-IL1-β, TNF, and IL-6, were strate that Vsig4 deficiency exacerbates macrophage-mediated −/− dramatically increased in virus-infected Vsig4 PEMs (Fig. 2f), inflammation, which deteriorates MHV-3 virus-induced FH. suggesting Vsig4 deficiency promotes macrophage-derived inflammation in vivo. Consistent with this, Vsig4 deficiency in VSIG4 attenuates LPS-induced macrophage activation in vitro. liver Kupffer cells also resulted in higher levels of these factors Our above data from in vivo experiments demonstrated that deposited in the infected liver tissues, as detected by qRT-PCR Vsig4 deficiency promotes macrophage-derived inflammation, we ab c d H&E staining TUNEL –/– –/– * Vsig4 WT Vsig4 WT ** 80 NS 60 0 0h 48h PI WT WT (n =5) * ** –/– 20 Vsig4 4000 1 (n =10) NS 0 123 4 0 300 –/– Vsig4 0 –/– Days post infection WT Vsig4 0 h 48 h PI –/– WT Vsig4 ef 8 2.0 6 0 h 12 h 0 h 12 h 0 h 12 h ** * * * 3 6 10.71 48.61 1.5 6.24 44.31 34.44 4.05 2 3 4 1.0 WT FSC 1 2 0.5 0 0 0 0 8.37 90.76 65.19 200 60 2.0 4.69 2.0 7.59 59.43 *** * * * –/– Vsig4 Live/death 150 45 1.5 1.5 100 1.0 1.0 F4/80 F4/80 F4/80 50 0.5 0.5 F4/80 0 0 0 0 –/– WT Vsig4 –/– –/– WT Vsig4 WT Vsig4 KD #1 #2 #3 #4 #1 #2 #3 #4 #1 #2 #3 #4 #1 #2 #3 #4 600 * 800 300 ** 4 NS 800 NS FGL2 3 600 400 NS 17 TNF * 31 Pro-IL-1β 300 2 400 400 NS NS * NS 200 NS 1 200 200 IL-1β p17 NS 21 IL-6 0 0 0 0 0 0 h 24 h 48 h 0 h 24 h 48 h 0 h 24 h 48 h 0 h 24 h 48 h 0 h 24 h 48 h IFN-γ –/– WT Vsig4 41.6 Actin-β 24 h PI 48 h PI −/− Fig. 2 Vsig4 deficiency exacerbates MHV-3-induced fulminant hepatitis. The Vsig4 mice and age-matched C57BL/6 WT littermates were infected with MHV-3 (100 PFU/mouse) via i.p. injection. a The survival was monitored. b H&E staining of liver, and TUNEL staining of cell apoptosis, scale bar = 20 μm, n = 5–8 per group, arrow indicated positive cells. c Serum ALT and AST levels at 0 h and 48 h post infection (PI), n = 5–8 per group. d Plaque assay of virus titers in livers at 48 h PI. e qRT-PCR analyzing proinflammatory cytokines in PEMs at 12 h and in liver tissues at 72 h of MHV-3 infection. f Flow cytometry analyzing TNF, pro-IL1-β, and IL-6 from PEMs after 12 h of virus infection. g Western blot analyzing proinflammatory cytokines in infected livers at 24 h and 48 h PI, n = 4 per group. h ELISA of serum concentration of proinflammatory mediators, n = 5–10 per group. Error bar, s.e.m. *p < 0.05, **p < 0.01, ***p < 0.001 and NS, p > 0.05. a was analyzed by log-rank test and others are calculated by Student’s t-test. Data are representative of six (a) and three (b–f, h) independent experiments 4 NATURE COMMUNICATIONS 8: 1322 DOI: 10.1038/s41467-017-01327-4 www.nature.com/naturecommunications | | | Livers % of survival PEMS Fgl2 mRNA relative Fgl2 mRNA relative Tnf mRNA relative Tnf mRNA relative 72 h 48 h 0 h Il-6 mRNA relative Il-6 mRNA relative Il-1 mRNA relative Il-1 mRNA relative FGL2 (pg/mL) SSC SSC SSC IFN-γ (ng/mL) Pro-IL-1β TNF (pg/mL) Serum AST (U/mL) Serum ALT (U/mL) TNF IL-1β (pg/mL) IL-6 IL-6 (pg/mL) Liver viral titers log (PFU/g) 10 NATURE COMMUNICATIONS | DOI: 10.1038/s41467-017-01327-4 ARTICLE then investigated the role of VSIG4 in regulating macrophage Previous work has shown that the complement C3b is the activation in response to LPS in vitro. PEMs that are isolated natural ligand of VSIG4 . To address whether VSIG4 regulates −/− from Vsig4 mice appeared to present with an abrupt surge of macrophage activation is dependent on C3b, we used lentiviral −/− M1-like proinflammatory gene transcripts, such as Il-1β and Tnf, vectors to overexpress VSIG4 in C3 BMDMs and subjected at the very early stage (1 h and 6 h) of LPS exposure (Fig. 3a). This these cells to LPS treatment (2 μg/ml). ELISA showed that the was validated by the existence of higher levels of IL-1β, TNF, and secretion of IL-6 and IL-1β was still dramatically downregulated −/− + −/− IL-6 protein in the supernatants of LPS-stimulated Vsig4 PEM in VSIG4 C3 BMDMs than their controls (Fig. 3h). These cultures compared to the WT counterparts (Fig. 3b). Western blot combined data suggest that VSIG4-mediated cytokine production also confirmed these results (Fig. 3c). In addition, surface in macrophages is C3b independent. expression of M1 activation markers, including B7-H1, B7-DC, −/− B7-H3, and CD40, was higher in LPS-treated Vsig4 PEMs (Fig. 3d). These results suggest that VSIG4 is important for VSIG4 reprograms pyruvate metabolism and mtROS genera- controlling macrophage activation. tion. To investigate the molecular mechanisms for VSIG4- To avoid the cellular heterogeneity of conventional PEMs, we mediated macrophage activation, we constructed a platform using next chose a macrophage line, RAW264.7 cells, as a homogeneous VSIG4 RAW264.7 cells to test the dependence of macrophage model to examine the functional specificity of VSIG4. RAW264.7 activation on VSIG4. Inspired by recent studies showing that cell cells are lack of Vsig4 transcription, but they with lentiviral- metabolism has an important function in supporting macrophage mediated restoration of Vsig4 expression (Len-Vsig4) exhibited a activation and polarization , we measured cell metabolism reduction in LPS-induced M1 gene (Il1b, Il6, and Tnf) transcripts thereafter. Figure 4a shows that VSIG4 did not affect LPS- compared to the control counterparts (Fig. 3e). Similarly, ELISA induced glucose uptake. However, it inhibited lactate, pyruvate data also showed that the levels of TNF, IL-6, IL-1β, and IL-12p40 and acelyl-CoA accumulation after 6 h of LPS administration. protein in the supernatants of the LPS-treated VSIG4 RAW264.7 The mitochondrial oxygen consumption rate (OCR) in the cultures were significantly lower than the controls (Fig. 3f). These macrophages was then investigated, although VSIG4 did not VSIG4 RAW264.7 cells were also incapable of mounting LPS- affect oxygen consumption under normal conditions (Supple- induced CD40 upregulation (Fig. 3g). These combined data imply mentary Fig. 6a, b), it appeared to drastically downregulate that VSIG4 inhibits LPS-induced macrophage activation in vitro. oxygen consumption after LPS exposure, both in basal and ab c –/– WT Vsig4 Il-1 Tnf 400 20 NS NS 400 KD 015 45 60 0 15 45 60 LPS (min) 4 12 NS pro-IL-1β 300 15 3 9 NS IL1β p17 * 17 200 10 NS 200 2 6 NS TNF NS NS * 3 100 100 1 NS NS NS IL-6 0 0 0 0 41.6 Actin-β 016 (h) 016 (h) 016 24 (h) 01624 (h) 016 24(h) –/– –/– WT Vsig4 WT Vsig4 de f NS 6 30 Il-1 Il-6 Tnf *** * 4 20 20 4 8 * * 2 NS 10 NS NS 15 3 10 2 0 1 624 (h) 0 12 6 4 (h) 4 Len-Cont. PD-L1 PD-L2 NS Len-Vsig4 NS 2 5 1 NS NS ** NS 150 60 0 0 ** NS 0 6 (h) 06 (h) 0 6 (h) 100 40 NS NS NS Len-cont. Len-Vsig4 50 20 B7-H3 CD40 0 0 –/– 016 24(h) 0 1 6( 24h) WT Vsig4 8000 40 0 h 6 h LPS (2μg/ml) * NS Len-Cont. p = 0.07 Len-cont. 4000 * Len-Vsig4 NS Len-Vsig4 NS NS 026 12(h) 0 6 12 (h) CD40 Fig. 3 VSIG4 impedes LPS-induced macrophage M1 polarization in vitro. PEMs were treated with LPS (2 μg/ml), a qRT-PCR analysis of Il-1β and Tnf transcripts. b ELISA of cytokines in cultured supernatants. c Western blot analyzing cytokine protein expression. d Flow cytometry analyzing surface expression of activation markers. RAW264.7 cells stably infected with lentiviral control vectors (Len-cont.) or vectors encoding Vsig4 (Len-Vsig4), cells were further treated with LPS (2 μg/ml), e qRT-PCR analysis of Il-1β, Il-6, and Tnf transcripts. f ELISA detecting cytokines in cultured supernatant. g Flow −/− cytometry analyzing surface expression of CD40. h C3 BMDMs were tranfected to overexpress VSIG4, and cells were further treated with LPS (2 μg/ ml), the secretion of IL-6 and IL-1β was detected by ELISA. Error bar, s.e.m. *p < 0.05, **p < 0.01, ***p < 0.001 and NS, p > 0.05 (Student’s t-test). Data are representative of three independent experiments NATURE COMMUNICATIONS 8: 1322 DOI: 10.1038/s41467-017-01327-4 www.nature.com/naturecommunications 5 | | | Count Count Count mRNA relative (folds) mRNA relative (folds) mRNA relative (folds) Il-1β (pg/ml) mRNA relative (folds) IL-6 (ng/ml) TNF (ng/ml) mRNA relative (folds) Il-6 (pg/ml) IL-1β (pg/ml) IL-1β (pg/ml) TNF (ng/ml) IL-12 p40 (ng/ml) IL-6 (ng/ml) ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-017-01327-4 maximal OCR (Fig. 4b, c), implying that VSIG4 inhibits mito- a substrate that in turn stimulates HIF-1α-dependent IL-1β chondrial oxidation during macrophage activation. expression . Nevertheless, LPS-induced succinate accumulation It is believed that glucose oxidation via the mitochondrial and HIF-1α upregulation in macrophages appear to be unaffected electron transport chain is a major source of mtROS upon cells by the presence of VSIG4 (Supplementary Fig. 7), suggesting undergoing aerobic metabolism . The fact that VSIG4 inhibited VSIG4-mediated macrophage activation is HIF-1α independent. mitochondrial oxidation led us to investigate the status of mtROS, which can induce macrophage M1 activation through activating NF-κB and stabilizing HIF-1α . Compared to the controls, flow VSIG4 enhances PDK2 expression in macrophages. The cytometry showed that VSIG4 RAW264.7 cells had a signifi- observation of VSIG4 inhibiting mitochondrial oxidation and cantly less mtROS secretion, especially in response to LPS mtROS secretion led us to investigate the underlying molecular stimulation (Fig. 4d). Conversely, LPS exposure caused increased mechanisms. Regulation of pyruvate metabolism largely relies on −/− mtROS secretion in Vsig4 PEMs in vitro (Fig. 4e). Moreover, PDH, whose activity is inhibited by PDKs via phosphorylation . −/− both ATMs from Vsig4 obese mice and PEMs isolated from In examination of the 4 Pdk isoform expressions in BMDMs and −/− MHV-3-infected Vsig4 mice had higher levels of mtROS than we found that the Pdk2 mRNA and protein levels were sig- −/− their WT littermates (Fig. 4f). Inhibition of mtROS production by nificantly lower in Vsig4 macrophages than that in their WT using diphenyliodonium chloride (DPI) can efficiently block LPS- counterparts (Fig. 5a, b). This low level of PDK2 was responsible induced IL-6 secretion from both RAW246.7 and VSIG4 for the appreciable decreases in phosphorylation of PDH (p- S300 S293 RAW264.7 cells (Fig. 4g). This implies that VSIG4 inhibits PDH-E1α and p-PDH-E1α ), as detected by western blot macrophage M1 activation mainly by reducing pyruvate oxida- and immunofluoresence staining (Fig. 5b, c). The absence of tion and mtROS generation. Vsig4 in Kupffer cells also resulted in PDK2 reduction and lower S300 S293 The other important consequence of LPS-induced metabolic p-PDH-E1α /E1α levels in the liver tissues, both in unin- reprogramming in macrophages is the accumulation of succinate, fected and at 48 h of MHV-3-infected conditions (Fig. 5d). 8 60 200 NS 6 150 12 NS 8 4 30 NS 100 Len-Cont. NS Len-Vsig4 2 15 50 4 0 0 0 0 bc d Oligo FCCP AA+AR Basal OCR Maximal OCR LPS (2μg/mL) Mock * * 120 160 * * Len-cont. 100 Len-Vsig4 90 120 60 80 015 35 55 75 0 0 mtROS Times (min) Isotype Len-cont. Len-Vsig4 Len-cont. Len-Vsig4 ef 0 h 6 h LPS (2 μg/ml) Obese MHV-3 12 h PI Mock LPS LPS+DPI mtROS mtROS Len-cont. Len-Vsig4 –/– –/– WT Vsig4 WT Vsig4 Fig. 4 VSIG4 reprograms pyruvate metabolism and mtROS generation. VSIG4 RAW264.7 cells and their controls were treated with LPS (2 μg/ml) for 6 h. a Colorimetric/Fluorometric assay of glucose uptake, pyruvate, Acelyl-CoA and Lactate levels, n = 5 per group. b OCR of LPS-treated RAW264.7 cells by Seahorse XFp assay. OCR detected before and after sequential treatment with ATP synthase inhibitor Oligo, mitochondrial uncoupling agent FCCP, ETC inhibitors AA+AR at indicated times, n = 5 per group. c OCR at basal and maximal levels of the indicated conditions was plotted in bar graphs. d mtROS −/− secretion was detected by flow cytometry. e Vsig4 PEMs and the WT controls were treated with LPS for 0 and 6 h, and mtROS secretion was detected by flow cytometry. f mtROS secretion from ATMs of obese mice and PEMs from 12 h of MHV-3-infected animals was compared by flow cytometry. g RAW264.7 cells were treated with mtROS inhibitor DPI (10 μM) for 48 h in advance, cells were then added with LPS (2 μg/ml) for an additional 6 h, IL-6 in the supernatant was detected by ELISA. Error bar, s.e.m. *p < 0.05 and NS, p > 0.05 (Student’s t-test). Data are representative of three independent experiments 6 NATURE COMMUNICATIONS 8: 1322 DOI: 10.1038/s41467-017-01327-4 www.nature.com/naturecommunications | | | Untreated LPS (2 μg/ml) Untreated LPS (2 μg/ml) Untreated LPS (2 μg/ml) Untreated LPS (2 μg/ml) Count OCR pmol/min Glucose uptake μg/μg protein Pyruvate (ng/μL) Count OCR pmol/min Acelyl-CoA nmol/μg protein OCR pmol/min Lactate (RU) Count IL-6 (ng/ml) NATURE COMMUNICATIONS | DOI: 10.1038/s41467-017-01327-4 ARTICLE Conversely, transient expression of VSIG4 in RAW264.7 cells with a significant reduction in IL-6/TNF secretion and limitation appeared to enhance PDK2 expression, and increase PDH- of CD40 upregulation (Fig. 5i, j). Collectively, these results S300 E1α phosphorylation, especially after LPS administration suggest that VSIG4 controls macrophage M1 activation by (Fig. 5e). Therefore, the PDH activity was decreased (Fig. 5f). regulating the PDK2-dependent pyruvate mitochondria meta- These data suggest that VSIG4 promotes PDK2 upregulation in bolic axis. macrophages. To validate the role of PDK2 in LPS-induced macrophage activation, Pdk2 expression was silenced in RAW264.7 cells by VSIG4 promotes DK2 by activating PI3K/Akt–STAT3.To using the specific shRNA (sh-Pdk2). Interestingly, knock-down of address the underlying mechanism of VSIG4 promotes PDK2 Pdk2 appeared to enhance oxygen consumption, both under upregulation in macrophages, we focused on the PI3K–Akt sig- normal conditions (Supplementary Fig. 6c, d) and after 2 h of LPS naling machinery in consideration of the fact that this pathway is administration (Fig. 5g), implying that PDK2 inhibits mitochon- essential for cellular metabolism, in addition to other functions drial oxidation. Moreover, enhancing mitochondrial oxidation such as cell growth, survival etc. . Western blotting data showed −/− caused the elevation of the basal mtROS production, especially that Vsig4 PEMs and BMDMs had obviously decreased in Akt after LPS stimulation (Fig. 5h), together with augmenting IL-6 phosphorylation (Fig. 6a). However, lentivirus-mediated over- and TNF secretion (Fig. 5i), as well as promoting CD40 expression of VSIG4 in RAW264.7 cells resulted in higher p- −/− ser473 expression (Fig. 5j). Similarly, LPS-treated Pdk2 BMDMs also Akt expression compared to mock-infected controls manifested with enhanced mtROS secretion, promoted the (Fig. 6b), suggesting that VSIG4 transfers a feedback signal, transcription of proinflammatory cytokine genes, as well as licensing macrophages for Akt activation. To identify the func- enhanced CD40 expression compared to WT counterparts tional motifs of VSIG4 that are responsible for augmenting Akt (Supplementary Fig. 8). Conversely, lentiviral overexpression of phosphorylation, we created a series of point and truncation PDK2 in RAW264.7 cells has opposite effects, not only by mutants of the molecule and tested their activity in affecting Akt ser473 quenching of the basal level but also by preventing the LPS- activation. The data from probing with p-Akt suggested that induced upregulation of mtROS (Fig. 5h). This was associated the c-terminal residues 267–280 aa of VSIG4 were critical for ac d e –/– WT WT Vsig4 –/– Vsig4 –/– KD WT Vsig4 NS 2 46 PDK2 NS NS #1 #2 #3 #4 #1 #2 #3 #4 KD S300 43 p-PDH-E1α 1 46 PDK2 s293 Actin-β 43 p-PDH-E1α 41.6 S300 43 p-PDH-E1α 015 45 LPS (min) 43 PDH b f 37 GAPDH s293 KD p-PDH-E1α 46 PDK2 S300 43 p-PDH-E1α 200 NS S300 43 p-PDH-E1α 43 PDH s293 43 p-PDH-E1α 37 GAPDH 43 PDH S300 p-PDEH-E1α /Mito tracker/DAPI 0 41.6 Actin-β 0 h 6 h (LPS) PEMs BMDMs Len-Cont. Len-Vsig4 hi j NS NS 0 h 6 h LPS (2 μg/ml) 150 Oligo FCCP AA+AR 1500 5 * * * * 1200 4 Cont. ** Sh Pdk2 900 3 sh-Scr 600 sh-Pdk2 NS NS 300 1 2 3 4 2 3 4 010 10 10 010 10 10 0 h 6 h 12 h (LPS) 0 h 6 h 12 h (LPS) 0 mtROS Red CD40 0 255075 100 sh-Scr sh-Pdk2 sh-Pdk2 sh-Scr Times (min) Basal OCR Maximal OCR *** 12 80 *** Len-Cont. 120 160 * * Len-Pdk2 *** *** NS 60 80 NS 3 20 2 3 4 2 3 4 010 10 10 010 10 10 30 40 0 CD40 mtROS Red 0 h 6 h 12 h (LPS) 0 h 6 h 12 h (LPS) 0 0 Len-Cont. sh-Pdk2 Len-Cont. Len-Pdk2 Len-Cont. Pdk2-shRNA −/− Fig. 5 VSIG4 triggers PDK2 expression in macrophages. Macrophages from WT and Vsig4 mice were collected. a qRT-PCR detection of 4 Pdk isoforms s300 s293 s300 in BMDMs. b Western blot analyzing PDK2, p-PDH-E1α , p-PDH-E1α , and total PDH. c The location of p-PDH-E1α in mitochondria was analyzed s300 s293 by immunofluoresence double staining, scale bar = 20μm. d Western blot of PDK2, p-PDH-E1α , p-PDH-E1α in liver tissues at 0 h and 48 h PI. RAW264.7 cells were transfected to expression of Vsig4, and cells were further treated with LPS (2 μg/ml), e Western blot analysis of PDK2 and p-PDH- s300 E1α . f PDH activity analysis, n = 6 per group. The expression of Pdk2 in RAW264.7 cells was silenced by shRNA or enhancing Pdk2 expression by lentivirus infection. g Seahorse analysis of OCR after 2 h of LPS treatment (up), and basal and maximal OCR of the indicated conditions was plotted in bar graphs (down), n = 5 per group. h Flow cytometric assay of mtROS secretion after LPS administration. i ELISA of IL-6 and TNF in cultured supernatants, n = 4 per group. j Flow cytometric assay of LPS-caused CD40 expression at 6 h. Error bar, s.e.m. *p < 0.05,**p < 0.01, ***p < 0.001 and NS, p > 0.05 (Student’s t-test). Data are representative of three independent experiments NATURE COMMUNICATIONS 8: 1322 DOI: 10.1038/s41467-017-01327-4 www.nature.com/naturecommunications 7 | | | Pdk1 Pdk2 WT –/– Pdk3 Vsig4 Pdk4 WT –/– Vsig4 Len-Cont. Len-Vsig4 Len-Cont. Len-Vsig4 Len-Cont. Len-Vsig4 OCR pmol/min OCR pmol/min mRNA relative (folds) OCR pmol/min PEMs BMDMs Count Count 48 h 0 h IL-6 (pg/ml) IL-6 (ng/ml) TNF (ng/ml) TNF (ng/ml) PDH activity Count Count ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-017-01327-4 mediating Akt phosphorylation (Fig. 6b). Several families of resulting in PDK2 downregulation (Fig. 6g). Conversely, silenced kinases phosphorylate both serine and threonine residues in Stat3 expression in VSIG4 RAW264.7 cells by specific shRNA target substrates, thus result in three dimensional changes of the also decreased basal PDK2 expression (Fig. 6h). Furthermore, protein structure and thereby alter its enzymatic activity or affects using ChIP-qPCR, we found that LPS induces p-STAT3 its ability to interact with other proteins . There are two serine recruitment to one of the two putative binding sites at the 273 276 270 residues (Ser and Ser ) and two threonine residues (Thr −1,298 bp but not the −2,934 bp of Pdk2 promoter region, and and Thr ) in the c-terminal 267–280 residues of VSIG4, and the present of VSIG4 signaling markedly promotes this recruit- 270 273 further assessment revealed that mutation of Thr , Ser , and ment (Fig. 6i). Taken together, these data show that VSIG4 276 274 Ser , but not Thr , to Ala respectively, could successfully induces PDK2 expression via activating the PI3K/ Ser473 inhibited LPS-caused Akt phosphorylation in RAW264.7 Akt–STAT3 signaling pathway. cells, thus prevented PDK2 upregulation (Fig. 6c), suggesting To address whether VSIG4 promotes PDK2 expression in 270 273 276 Thr , Ser , and Ser residues of VSIG4 play essential role in macrophage is dependent on C3b, the human monocyte cell line, mediating PDK2 expression. The direct involvement of Akt in THP-1 cells, was transfected to overexpress human VSIG4, and modulating PDK2 expression was demonstrated by showing that these VSIG4 THP-1 cells were further induced to be macro- treating VSIG4 Raw264.7 cells with the Akt inhibitor MK-2206 phages by PMA stimulation. Additionally, cells were activated (Fig. 6d) or the PI3K inhibitor Ly294002 (Fig. 6e), all resulted in with microbeads-conjugated C3b, and western blot indicates that typical downregulation of p-Akt and PDK2. These combined data microbeads-C3b does not affect the basal and LPS-induced PDK2 −/− suggest that VSIG4 triggers PDK2 upregulation by activating the expression (Fig. 6j). Furthermore, overexpress VSIG4 in C3 PI3K/Akt pathway. BMDMs also increased basal and LPS-induced PDK2 expression Chromatin immunoprecipitation and massive parallel sequen- (Fig. 6j). These combined data suggest that VSIG4-mediated cing (ChIP-Seq) have demonstrated that the Pdk2 promoter PDK2 upregulation in macrophages is C3b independent. region has two binding sites for the signal transduction and activator of transcription-3 (STAT3) , which provides a basis for analyzing the signaling pathways responsible for VSIG4-PI3K/ Promoter methylation inhibits Vsig4 gene transcription. Our Akt-induced PDK2 upregulation. Interestingly, we found that data indicate that VSIG4 suppresses macrophage-dependent both the PI3K inhibitor Ly294002 (Fig. 6e), and the Akt inhibitor inflammation by augmenting PDK2 expression, which high- MK-2206 (Fig. 6f), could successfully inhibit LPS-induced STAT3 lights a plausible therapeutic intervention for inflammatory dis- phosphorylation (p-STAT3) in VSIG4 RAW264.7 cells. Simi- orders through enhancing VSIG4 signaling. However, in larly, the STAT3 inhibitor S3I-201 was also able to impair LPS- agreement with the previous report , we found that the PEMs induced p-STAT3 expression in VSIG4 RAW264.7 cells, thus and liver tissues isolated from MHV-3-infected mice manifested ab c SP IgV TM CT KD KD ser473 60 p-Akt ser473 Vsig4 60 p-Akt KD Δ119–280 ser473 46 PDK2 60 p-Akt 60 Akt Δ257–280 41.6 Actin-β 41.6 Actin-β 41.6 Actin-β Δ267–280 PEMS BMDMs 0h 1h LPS (2 μg/ml) – + + + + + LPS (2 μg/ml) de f – – – + LPS (2 μg/ml) – + + ++ LPS (2 μg/ml) – – + + LPS (2 μg/ml) – + – +++ LPS (2 μg/ml) – + – + Ly294002 (10 μM) KD 00 1 5 10 MK-2206 (μM) KD KD – + – + S3I-201 (100 μM) – 0.5 MK-2206 (μM) ser473 – 15 10 ser473 80 p-Akt KD 60 p-Akt 80 p-STAT3 80 STAT3 Ser307 80 p-STAT3 60 Akt STAT3 80 p-STAT3 46 PDK2 46 PDK2 46 PDK2 41.6 Actin-β Actin-β 41.6 Actin-β 41.6 41.6 Actin-β hi –/– C3 BMDMs THP-1 cells 1.5 * ––– +++ C3b (20 μg/ml) Len-cont.–RAW264.7 mouse IgG *** KD 1.0 0 h 1 h 3 h 0 h1 h3 h LPS (2 μg/ml) KD Len-cont.–RAW264.7 p-STAT3 mAb KD 46 PDK2 * NS + 46 PDK2 46 PDK2 Vsig4 RAW264.7 mouse IgG 0.5 Actin-β NS NS + 41.6 41.6 Actin-β 41.6 Actin-β Vsig4 RAW264.7 p-STAT3 mAb – – + + LPS (2 μg/ml) –1298 –2934 ser473 Fig. 6 VSIG4 promotes PDK2 expression through activating PI3K/Akt–STAT3. a Western blot of Akt and p-Akt expression. b RAW264.7 cells were ser473 infected with different Vsig4 deletion constructs, cells were further treated with LPS (2 μg/ml), the expression of p-Akt was analyzed by western blot. ser473 + 273 276 270 274 + c Western blot of p-Akt and PDK2 in VSIG4 RAW264.7 cells with mutation of Ser , Ser , Thr , and Thr to Ala. VSIG4 RAW264.7 cells ser473 were treated with d the Akt inhibitor MK-2206, e the PI3K inhibitor Ly294002, and the expression of Akt, p-Akt , PDK2, and p-STAT3 was detected by western blot. f VSIG4 RAW264.7 cells were treated with the MK-2206, and the expression of STAT3/p-STAT3 was analyzed by western blot. g Western blot of p-STAT3/STAT3 and PDK2 in LPS-activated VSIG4 RAW264.7 cells followed with STAT3 inhibitor, S3I-201 (100 μM) treatment for 24 h. h Western blot of PDK2 in Stat3 silenced VSIG4 RAW264.7 cells. i The enrichment of p-STAT3 in Pdk2 gene promoter region was detected by ChIP-qPCR assay. j Human VSIG4 THP-1 cells were treated with microbeads-C3b (20 μg/ml) in the presence of LPS (2 μg/ml), and the expression of PDK2 was −/− detected by western blot. Moreover, C3 BMDMs were transfected to overexpress VSIG4, and cells were further treated with LPS (2 μg/ml) for an additional 3 h, and the expression of PDK2 was detected by western blot. Error bar, s.e.m. *p < 0.05, ***p < 0.0001 and NS, p > 0.05 (Student’s t-test). Data are representative of three independent experiments 8 NATURE COMMUNICATIONS 8: 1322 DOI: 10.1038/s41467-017-01327-4 www.nature.com/naturecommunications | | | WT –/– Vsig4 WT –/– Vsig4 Sh-Scr Sh-Stat3 #1 Sh-Stat3 #2 Mock Vsig4Δ119–280 Vsig4Δ257–280 Vsig4Δ267–280 Vsig4 Mock Vsig4Δ119–280 Vsig4Δ257–280 Vsig4Δ267–280 Vsig4 Vsig4 Thr →Ala Ser →Ala Thr →Ala Ser →Ala Vsig4 Len-Cont. Len-Vsig4 Len-Cont. Len-Vsig4 Binding (%) NATURE COMMUNICATIONS | DOI: 10.1038/s41467-017-01327-4 ARTICLE with lower VSIG4 expression as compared to the uninfected negatively regulates Vsig4 gene transcription (Fig. 7f). In controls (Fig. 7a, b). Moreover, administration of proin- accordance, proinflammatory stimuli seemed to have no addi- flammatory mediators (including LPS, TNF, MALP2, IFN-γ, poly tional inhibition on the promoter activity in RAW264.7 cells (I:C) or CpG) apparently can induce a transient sharp decline of transfected with M.SssI-methylated constructs (Fig. 7f). These Vsig4 gene transcription and protein in ex vivo PEMs (Fig. 7c, d). data demonstrate that Dnmt3a controls Vsig4 gene repression Thus, rapid VSIG4 downregulation appears to be a common through fast methylation of Vsig4 gene promoter. response of macrophages upon inflammatory stimulations. We also analyzed the genomic DNA sequences of isolated The epigenetic mechanisms, especially DNA methylation of BMDMs that were treated with various proinflammatory CpG sites within promoter regions, have recently been described mediators for 12 h. Proinflammatory stimuli all appeared to to mediate gene silencing . There are three types of mammalian induce very high incidence (98–100%) of methylation at a CpG DNA methyltransferases (Dnmt), distinct from Dnmt1 that is site (-374 bp) in the promoter region of Vsig4 gene, which is responsible for copying DNA methylation patterns during significantly elevated from a 75% basal methylation at this site in replication, Dnmt3a and Dnmt3b are important in de novo the untreated controls as detected by using the Sequenom DNA methylation . We thereafter hypothesized that inflamma- MassARRAY platform (Supplementary Fig. 10a, b, Supplemen- tion leads to silencing Vsig4 gene transcription through triggering tary Table 1). Furthermore, ChIP-qPCR assays reveal a significant the transcriptional activation of Dnmts. Western blot showed that enrichment of Dnmt3a in −374 bp of Vsig4 promoter region after the expression of Dnmt3a was upregulated in BMDMs in the BMDMs were treated with proinflammatory factors (Supple- response to proinflammatory stimuli (Fig. 7e), nevertheless, the mentary Fig. 10c). These combined data suggest that CpG at -374 expression of Dnmt1 and Dnmt3b appeared to be not affected bp site, probably with other CpG islands in the Vsig4 promoter under such conditions (Supplementary Fig. 9). However, 5-aza- region, play an essential role for negative feedback control of 2ʹ-deoxycytidine (AZAdC), a general DNA methyltransferases macrophage activation during inflammatory response. inhibitor, was able to effectively downregulate Dnmt3a expres- sion, and in agreement, macrophages treated with AZAdC appeared to be resistant to the proinflammatory factors-caused Forced overexpression of Vsig4 improves MHV-3-caused FH. VSIG4 downregulation (Fig. 7e). We therefore tried to transiently force the expression of Vsig4 in To validate the importance of methylation in controlling Vsig4 MHV-3-susceptible C57BL/6 WT mice using lentiviral vectors gene transcription, a 840 bp fragment of Vsig4 gene promoter in vivo. These mice expressed significantly higher levels of VSIG4 (−840/ + 1) was constructed into a luciferase reporter pGL3-basic in the livers on day 6 of transduction compared to control vector. The pGL3-Basic and Vsig4 promoter constructs (−840/ + infected animals (Fig. 8a). Interestingly, Vsig4-transgenetic mice 1) were further fully methylated by CpG Methyltransferase (M. had significant enhancing PDK2 expression while lessening PDH S300 SssI). We transfected RAW264.7 cells with these plasmids and phosphorylation (p-PDH-E1α ) in liver tissues at 72 h of compared their reporter luciferase activities. Interestingly, the MHV-3 infection (Fig. 8b), leading to lower levels of FGL2, TNF, luciferase activity of the M.SssI-methylated construct exhibited IL-1β, and IL-6 deposition in liver tissues (Fig. 8b), together with 53% reduction in promoter activity, indicating DNA methylation reducing liver damage (Fig. 8c), along with a considerably ab c d PBS TNF 1.5 Poly(I:C) LPS 12 h PI 0 h 48 h PI IFNγ LPS+IFN-γ MALP-2 CpG KD KD #1 #2 #1 #2 1.0 0 h PI 45 VSIG4 45 VSIG4 41.6 Actin-β 41.6 Actin-β 0.5 0 1 2 3 4 ** 10 10 10 10 10 ** VSIG4 ef pGL3-basic pGL3-Vsig4 promoter (–860/+1) 1.5 NS NS NS * Unmethylated KD NS 1.0 M.Sssl methylated NS 101.5 Dnmt3a NS NS 45 VSIG4 0.5 41.6 Actin-β Fig. 7 Vsig4 gene transcription is repressed Dnmt3a-mediated DNA methylation. The C57BL/6 WT mice were infected with MHV-3 (100 PFU/mouse), a the expression of VSIG4 on PEMs at 0 h and 12 h PI was detected by flow cytometry. b VSIG4 protein level in liver tissues was analyzed by western blot. The BMDMs were treated with proinflammatory factors for 12 h. c Vsig4 gene transcription was detected by qRT-PCR. d VSIG4 protein expression was evaluated by western blot. e The BMDMs were treated with Dnmts inhibitor-AZAdC (10 μM) for 72 h in advance, cells were then further added with proinflammatory mediators for 12 h, the expression of Dnmt3a and VSIG4 was assessed by western blotting. f Luciferase activity of the lysates from RAW264.7 cells transfected with unmethylated or M.SssI methylated pGL3-basic vector and the -840/+1 Vsig4 promoter constructs. Error bar, s.e.m. *p < 0.05, **p < 0.01 and NS, p > 0.05 (Student’s t-test). Data are representative of three independent experiments NATURE COMMUNICATIONS 8: 1322 DOI: 10.1038/s41467-017-01327-4 www.nature.com/naturecommunications 9 | | | PBS LPS TNF MALP-2 Poly(I:C) IFN-γ CpG PBS AZAdC AZAdC+LPS AZAdC+TNF AZAdC+MALP-2 AZAdC+Poly(I:C) AZAdC+IFN-γ AZAdC+CpG PBS Poly(I:C) IFN-γ MALP-2 TNF LPS+IFN-γ LPS CpG Untreated Untreated Untreated Untreated LPS TNF MALP-2 Poly(I:C) IFN-γ CpG Count Vsig4 mRNA suppression (suppression) Relative luciferase activity (folds) ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-017-01327-4 improved survival rate (Fig. 8d). These combined data suggest neutrophils is due to reduction in cellular nicotinamide adenine that increasing the expression of VSIG4 might have therapeutic dinucleotide phosphate oxidases activity that results in produc- potentials for fulminant hepatitis and other macrophage- tion of superoxide during the respiratory burst. On the other associated inflammatory disorders (Supplementary Fig. 11). hand, over production of ROS actively participates in the pathogenesis of inflammatory diseases including rheumatoid arthritis, multiple sclerosis, and thyroiditis through activating the inflammatory signaling pathways including mitogen-activated Discussion Macrophage activation relies on metabolic adaptation in response protein kinases (MAPK), NF-κB, and guanylate cyclase . ROS comes from various sources, such as peroxisomes, ubiquinone, to the surrounding micro-environmental stimuli. Macrophage plasticity determines its biological functions in immunity, activities of cytosolic enzymes and uncoupled nitric oxide syn- thases. However, recent data identify that mitochondria are a inflammation, and tissue homeostasis. Defining the mechanisms major source of physiological intracellular ROS that drives regulating macrophage metabolic patterns is critical for under- inflammation . mtROS can be sensed by the NLRP3 inflam- standing the pathology of inflammatory disorders and developing masome, resulting in caspase-1 activation and IL-1β matura- therapeutic interventions . We here demonstrate that VSIG4, a tion . Therefore, high levels of mtROS in vivo would impair B7 family-related protein that is expressed specifically in resting glucose metabolism and insulin sensitivity , and it probably macrophages, is able to inhibit macrophage activation by repro- explains why exaggerated macrophage activation promotes gramming mitochondrial pyruvate oxidation. Vsig4 deficiency pathogenesis of MHV-3-mediated hepatitis . Alternatively, apparently sways macrophage towards activation upon LPS mtROS also activates NF-κB and stabilizes HIF-1α, by thus it exposure in vitro. Conversely, overexpression of Vsig4 suppresses increases the activities of these two transcriptional factors leading M1 gene expression and reduces LPS-induced pyruvate oxidation and mtROS secretion by RAW264.7 cells. Interestingly, Vsig4 to upregulation of macrophage M1 genes . Here, we show that VSIG4 can inhibit pyruvate/acetyl-CoA conversation in deficiency affects the outcomes of inflammatory disorders in −/− animal models. For instance, Vsig4 mice are more susceptible RAW264.7 cells, leading to limitation of oxygen consumption (Fig. 4a–d). Interestingly, decreasing mitochondrial oxidation also to develop HFD-induced obesity and insulin resistance. Fur- −/− leads to the inhibition of LPS-induced mtROS secretion, along thermore, Vsig4 mice exhibit markedly higher mortality over with restriction of Il1b, Il6, and Tnf gene transcription and CD40 MHV-3 viral infection, clearly due to the exacerbated upregulation (Fig. 3). Moreover, PEMs isolated from MHV-3- macrophage-dependent inflammation in vivo. Finally and the −/− −/− infected Vsig4 mice or ATMs collected from HFD-fed Vsig4 most notably, the anti-inflammatory function of VSIG4 has been obese animals all exhibited with more mtROS secretion in vivo as validated by forced overexpression of VSIG4 in vivo, for which we compared to WT littermates (Fig. 4f). Similar to previous stu- show that excessive epigenetic VSIG4 dampens liver tissue dies , inhibition of mtROS activity by DPI successfully prevented inflammation and protects the susceptible mice from MHV-3 virus-induced FH. These studies demonstrate that VSIG4 inhibits LPS-mediated IL-6 production (Fig. 4g). These data suggest that VSIG4 inhibits macrophage activation via posing restriction on macrophage M1-associated inflammatory pathogenesis. It has long been known that ROS plays essential roles in mitochondrial oxidation and mtROS secretion. mtROS derives from pyruvate metabolism during OXPHOS in immune responses of macrophages and neutrophils to pathogens. The bacteria killing capability of activated macrophage and the mitochondria, in which the conversion of cytosolic pyruvate ab Len-Cont. Len-Vsig4 KD #1 #2 #3 #4 #1 #2 #3 #4 2.0 46 PDK2 1.5 S300 43 p-PDH-E1α 1.0 43 PDH 0.5 70 FGL2 0 21 IL-6 17 TNF 17 IL-1β p17 41.6 Actin-β c d Len-Vsig4 100 Len-Cont. Len-Vsig4 (n =8) Len-Cont. (n =8) 0 2 4 6 8 101214161820 Days post infection Fig. 8 Forced overexpression of Vsig4 improves MHV-3-induced hepatitis. C57BL/6 WT mice were infected with lentivirus (10 PFU/mouse) to induce the expression of Vsig4 in vivo, these mice were further infected with MHV-3 at day 6. a Liver Vsig4 gene transcription was analyzed by qRT-PCR at day 6, n = s300 5 per group. b Western blotting for PDK2, p-PDH-E1α , FGL2, and proinflammatory cytokines TNF, IL-6 and IL-1β in liver tissues at 72 h of MHV-3 infection, n = 4 per group. c The architecture of the liver tissues at 72 h of infection was compared by H&E staining, scale bar = 20 μm, n = 5 per group. d The survival was monitored for a total of 20 days. Error bar, s.e.m. a *p < 0.05 was analyzed by Student’s t-test, and d was analyzed by log-rank test. Data are representative of three independent experiments 10 NATURE COMMUNICATIONS 8: 1322 DOI: 10.1038/s41467-017-01327-4 www.nature.com/naturecommunications | | | Len-Cont. Len-Vsig4 mRNA relative (folds) % of survival NATURE COMMUNICATIONS | DOI: 10.1038/s41467-017-01327-4 ARTICLE into mitochondrial acetyl-CoA is partially regulated by the contributes to limiting macrophage M1-associated inflammation, activity of PDH . PDH activity is enhanced via its depho- and its deficiency can cause over inflammatory disorders. sphorylation by phosphopyruvate dehydrogenase phosphatase Identification of mediators that regulate Vsig4 expression will (PDP), and inducing PDP1 expression has been shown to pro- bring insights into understanding the inflammatory response in mote mitochondrial OXPHOS, mtROS production and M1 gene the host. The epigenetic mechanisms, such as DNA methylation expression . Conversely, PDK-mediated phosphorylation of the and histone modification, have recently been shown to mediate 232 293 53 PDH-E1α subunit at 3 serine sites, including Ser , Ser , and gene silencing . DNA methylation is a covalent modifications, in 300 42 Ser , appeared to cause PDH suppression . Recent studies have which DNA methyltransferase catalyze cytosine methylation shown that PDK/PDH axis essentially controls cytokine secretion using S-adenosyl methionine as a methyl donor . It has been in macrophages. For example, overexpression of PDK2 inhibits reported that the expression of tumor suppressor genes in several radiation-induced cytokine expression , whereas macrophages human cancers were inhibited upon DNA methylation at their 37, 55 from Pdk2/4-deficient animals appear to produce less IL-1β and CpG sites within the promoters . We here found that Vsig4 TNF in association with a low level of lactic acid . We here mRNA and protein were decreased dramatically in PEMs that identify that the VSIG4-dependent inhibition of macrophage M1 were treated with inflammatory factors in vitro (Fig. 7c–e). activation relies on PDK2. LPS stimulation resulted in augmen- Conversely, blocking Dnmt3a activity by the inhibitor, 5-aza-2ʹ- tation of mtROS secretion and inflammatory gene expression in deoxycytidine, could effectively prevent proinflammatory factors Pdk2 silenced RAW264.7 cells and Pdk2 deficient BMDMs induced Vsig4 downregulation (Fig. 7e). Further studies demon- (Fig. 5h–j, Supplementary Fig. 8). In agreement, overexpression of strated that the luciferase activity of the M.SssI-methylated Vsig4 Pdk2 results in opposite effects (Fig. 5h–j). Along with previous promoter (-840/+1) was reduced by 53% comparing to the work suggesting that PDK1 is involved in promoting aerobic unmethylated controls, and proinflammatory stimuli were unable glycolysis in macrophages , these combined data highlight a to further inhibit the promoter activity in RAW264.7 cells signaling pathway that governs the fate of macrophages differ- transfected with the M.SssI-methylated Vsig4 promoter con- entiation and function and Pdk status is involved in mediating structs (Fig. 7f). We also found that some CpG sites in Vsig4 metabolic programs for balancing between glycolysis and glucose promoter region (like CpG at -374 bp site) were methylated by oxidation. proinflammatory stimuli as detected by using the Sequenom Substantial evidence demonstrates that obesity is a chronic MassARRAY platform (Supplementary Fig. 10, Supplementary low-grade inflammatory disease . In obesity, adipocytes can Table 1). Together, these data suggest that proinflammatory release proinflammatory mediator like CC chemokine ligand factors repress Vsig4 gene transcription through inducing (CCL)-2, and monocyte chemoattractant molecule (MCP)-1, Dnmt3a activity, causing fast methylation of Vsig4 promoter which induce the recruitment of ATMs . ATM secretes proin- region, for which CpG at -374 bp site, probably with other CpG flammatory cytokines and forms the inflammatory circuit which islands within the Vsig4 promoter region, plays a critical role in blocks the insulin action in adipocytes and leads to insulin regulating Vsig4 transcription. resistance . Here we demonstrate that ATMs isolated from HFD- Recent studies have illustrated that VSIG4 suppresses T-cell −/− fed Vsig4 obese mice have higher levels of proinflammatory cytokine production and causes cell cycle arrest. Injection of factors such as TNF, IFN-γ, and IL-1β (Fig. 1i–k). It was eluci- soluble VSIG4-Ig protein causes a reduction in IFN-γ production dated that proinflammatory cytokines promote obesity-associated by antigen-specific CD8 T cells and down-modulates Th- insulin resistance through activating inflammatory signaling dependent IgG responses in vivo . Additionally, administration pathways, including the stress-responsive c-Jun NH2-terminal of VSIG4-Ig protein prolonged mouse survival in a ConA- kinase (JNK1/2) and AMP-activated protein kinase α2 induced hepatitis model and protected against the pathogenesis of 25, 56 (AMPKα2), inhibitor of κB kinase, and extracellular signal- many inflammatory diseases . One of the explanations for this regulated kinase 1/2 (ERK1/2) as well as MAPKp38, collectively effect is that VSIG4 binds to an as-yet-unidentified inhibitory causing inhibition of serine/threonine phosphorylation of the receptor/ligand on T cells. Alternatively, the surface VSIG4 on 48–50 docking protein IRS-1 . In agreement with previous studies, macrophages and CR1/CR3 on T cells bind to the same multi- −/− we found that Vsig4 obese mice exhibit exacerbated insulin meric C3b or iC3b molecules, thus triggering signals to suppress ser473 57 resistance along with diminished p-IRS-1 and p-Akt in the T-cell activation . Here, we showed that the secretion of cyto- WAT and the liver tissues (Fig. 1h), indicating that Vsig4 defi- kines in T cells that isolated from liver tissues of 72 h post MHV- −/− ciency eliminates the negative control signals, thus allowing 3-infected Vsig4 animals were similar to their WT counter- ATMs to produce proinflammatory cytokines and switching on parts (Supplementary Fig. 3), implicating VSIG4 does not affect HFD-associated insulin tolerance in vivo. T-cell activation in our models. Conversely, we demonstrate that Microorganism infection and tissue injuries trigger the VSIG4 delivers negative feedback signals to macrophages, recruitment of inflammatory macrophages from the circulation resulting in PI3K–Akt–STAT3 dependent PDK2 upregulation, into the affected tissues . The MHV-3 virus provokes a mouse and finally impedes mtROS-dependent M1 macrophage activa- strain-associated severe liver disease that has been used as a tion, suggesting VSIG4 negatively controls macrophage activa- model for investigating human viral fulminant hepatitis. This tion. However, microbeads-C3b did not affect the expression of exacerbation of macrophage-dependent cytokine storm directly PDK2 from LPS-activated VSIG4 THP-1 cells. Additionally, the causes hepatic necrosis and induces lethality in susceptible mouse expression of PDK2 and the secretion of cytokines like IL-6 as 7 −/− strains . In support, normal induction of macrophage apoptosis well as IL1-β were still downregulated in LPS-stimulated C3 can appropriately control local liver tissue fibrinogen deposition BMDMs (Figs. 3h, 6j), these combined data suggest that inhibi- and local tissue injuries in this fulminant hepatitis model .As tion of macrophage activation by VSIG4 is C3b independent. expected, we have found that MHV-3 infection super-induces Further studies are needed to identify the potential interacting −/− production of a panel of inflammatory cytokines from Vsig4 partners for VSIG4. macrophages, and the virus-infected mice exhibit with more In summary, we have demonstrated that VSIG4 down- severe liver necrosis and high mortality (Fig. 2). In contrast, modulates macrophage activation and M1 polarization in forced expression of Vsig4 in susceptible mice provides significant response to inflammatory stimuli in vitro and in vivo. Mechan- protection against MHV-3-induced fulminant hepatitis (Fig. 8). istically, VSIG4 is capable of sending feedback signals in mac- These combined results demonstrate that VSIG4 functionally rophages to activate the PI3K–Akt–STAT3 signaling axis, leading NATURE COMMUNICATIONS 8: 1322 DOI: 10.1038/s41467-017-01327-4 www.nature.com/naturecommunications 11 | | | ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-017-01327-4 to PDK2 upregulation and activation, thus inhibiting mitochon- pyruvate, 2 mM glutamine, pH 7.4) and incubated in a non-CO incubator for 1 h. Four baseline measurements were taken before sequential injection of mitochon- drial pyruvate metabolism, suppressing mtROS secretion and drial inhibitors oligomycin, FCCP, and antimycin (AA) plus rotenone (AR) pro- M1-like gene expression through inducing PDH phosphorylation. vided by the manufacturer (#101706-100, Agilent Technologies). OCR was Therefore, we speculate that enhancing VSIG4 signaling may automatically calculated using the Seahorse XFp software. Every point represents result in beneficial effects on treating inflammatory disorders. an average of three different wells. Methods Immunohistochemistry and immunofluorescence double staining. Paraffin- −/− embedded tissue blocks were cut into 2.5 μm slices and were mounted on Mice. The complement C3 deficient (C3 ) mice (#003641) and the C57BL/6 −/− polylysine-charged glass slides. Endogenous peroxidase activity was blocked by mice were purchased from Jackson Laboratory. The Vsig4 mice were kindly exposure to 3.0% H O for 30 min. Antigen retrieval was performed in a citrate provided by Dr. M. van Lookeren Campagne (Department of Immunology, Gen- 2 2 −/− buffer (pH 6.0) at 120°C for 10 min. Sections were then incubated at 4 °C overnight entech, CA, USA). The Pdk2 mice were provided by Dr. C.R. Harris (Rutgers with anti-mouse FGL2 (#sc-100276, Santa Cruz, 1:100, mouse), anti-Fibrinogen Cancer Institute of New Jersey, USA). All mice were backcrossed ten times onto the (#ab118533, Abcam, Cambridge, England, 1:1000, Rabbit), anti-pro-IL-1β (#12242, B6 background to avoid unpredictable confounders. Specific pathogen-free male Cell Signaling Technology (CST), 1:100, mouse), anti-TNF (#3707, CST, 1:100, and age-matched mice (8–12 weeks old) were used for all experiments. Mice were rabbit), anti-IL-6 (#sc-130326, Santa Cruz, 1:200, mouse), and anti-IFN-γ (#sc- maintained in micro-isolator cages, fed with standard laboratory chow diet and 52557, Santa Cruz, 1:200, rat). After washing, the sections were incubated with the water, and housed in the animal colony at the animal center of the Third Military corresponding secondary antibodies for 2 h at room temperature. The Vecta-stain Medical University (TMMU). All animals received humane care according to the ABC kit (Vector Laboratories, San Diego, CA, USA) was used to perform the criteria outlined in the “Guide for the Care and Use of Laboratory Animals” avidin–biotin complex method according to the manufacturer’s instructions. Sec- prepared by the National Academy of Sciences and published by the National tions incubated with isotype and concentration matched immunoglobulins without Institutes of Health (NIH publication 86–23 revised 1985). All of the in vivo primary antibodies were used as isotype controls. Peroxidase activity was visualized experiments comply with the animal study protocol approved by the ethics com- with the DAB Elite kit (K3465, DAKO), and brown coloration of tissues repre- mittee of TMMU. sented positive staining. S300 To detect p-PDH-E1α and mitochondria co-localization, both BMDMs and Cells. The mouse macrophage cell line RAW264.7, human monocyte cell line PEMs were fixed with 4% paraformaldehyde, permeabilized with 0.1% Saponin in THP-1 and 293T cells were provided by the Cell Institute of the Chinese Academy PBS for 5 min, and blocked with PBS containing 2% BSA for 1 h at 4°C. The cells of Sciences (Shanghai, China). Mouse 17 clone 1 (17CL1) cells were purchased S300 were then stained with rabbit anti-p-PDH-E1α antibodies (#AP1046, 1:100, from ATCC. The cells were cultured in 6-well plates and propagated in DMEM Merk, Temecula, CA, USA) overnight at 4 °C, and then stained with Alexa supplemented with 10% FBS, 100 U/ml penicillin, and 100 μg/ml streptomycin. Fluor488-conjugated donkey anti-rabbit IgG (H+L) highly cross-adsorbed Peritoneal exudative macrophages (PEMs) were harvested and BMDMs induced by secondary antibody (#A-21206, Thermo Scientific, Billerica, MA, USA) for 1 h. M-CSF (#400-28, Peprotech, Rocky Hill, NJ, USA) as described previously . Finally, the sections were incubated with 1 μg/ml DAPI and MitoTracker Red (#M- 7512, Thermo Scientific) at 500 nM for 30 min at 37 °C. The results were analyzed using fluorescence microscopy (Zeiss Axioplan 2). Virus and infection. MHV-3 viruses were expanded in 17CL1 cells to a con- centration of 1 × 10 plaque forming units (PFU)/ml. Mice were received MHV-3 (100 PFU/mouse) via intraperitoneally (i.p.) injection. In some experiments, ELISA and western blotting. The concentration of cytokines, free fatty acid, and C57BL/6 WT mice were infected with lentivirus (1 × 10 PFU/mouse) via intra- triglycerides in the serum or the culture supernatants, the levels of succinate, venous injection to transiently force the expression of Vsig4 in vivo, and these mice pyruvate, acetyl-CoA, triglyceride and PDH enzyme activity in macrophages were were further infected with MHV-3 after 6 days. The virus-infected mice were measured by the ELISA according to the manufacturer’ introductions. ELISA Kits, euthanized on the indicated days. Liver damage was compared by H&E staining including TNF (#EK0527), IL-6 (#EK0411), IL-1β (#EK0394), and IL-12 p40 and cell apoptosis was measured using the Terminal Transferase dUTP Nick End (#EK0932) were from Boster Ltd. (Wuhan, China). FGL2 ELISA Kit was from Uscn Labeling (TUNEL) staining method according to the manufacturer’s instructions Life Science (#SEA512Mu, Wuhan, China). Insulin ELISA Kit was from Millipore (#12156792910, Roche, Mannheim, Germany). The virus titers in liver tissues were (#2617704, Billerica, MA, USA). Pico ProbeAcetyl-CoAFluorometric Assay kit was determined by plaque assay method. Briefly, the supernatant of liver tissue from BioVision (#K317-100, Milpitas, CA, USA). The Free Fatty Acid Quantitation homogenate was ten times step diluted. Mouse 17CL1 cells were seeded in 12-well Kit (#MAK044-1KT), PDH Activity Assay Kit (#MAK183-1KT), Pyruvate Assay plates, when reaches 80% fusion, cells were added with the diluted liver extraction Kit (#MAK071-1KT), Lactate Assay Kit (#MAK064-1KT), Succinate Colorimetric and incubated for 30 min under 37 °C, 5% CO condition. Cells were then added Assay Kit (#MAK184-1KT), and other chemicals were all from Sigma-Aldrich (St. with 1 ml 2% MethyCellulose DMEM medium (100 U/ml penicillin and 100 μg/ml Louis, MO, USA). streptomycin) per well, and then further incubated for 4 days. Cells added with the The expression of GAPDH (#2118, CST, Danvers, MA, USA,1:1000, rabbit), supernatant from uninfected liver tissue homogenate were used as negative con- 307 PDK2 (#sc-14486, Santa Cruz, 1:500; goat), p-IRS-1(p-Ser , #2381, CST, 1:1000, trols, whereas cells treated with the purified MHV-3 virus were used as positive 308 473 rabbit), Akt (#2920, CST, 1:1000, mouse), p-Akt (p-Thy , #4056; p-Ser , #4051, controls. Finally, cells were fixed with 4% paraformaldehyde and the viral titers 293 CST, 1:1000, rabbit), PDH (#3025, CST, 1:1000, rabbit), p-PDH-E1α (#AP1062, were determined by crystal violet staining assay. S300 Merk, 1:2000, rabbit), p-PDH-E1α (#AP1046, Merk, 1:2000, rabbit), STAT3 (#9139, CST, 1:1000, mouse), p-STAT3 (p-Tyr , #9145, CST, 1:2000, rabbit), −/− FGL2 (#sc-100276, Santa Cruz, 1:500, mouse), TNF (#3707, CST, 1:1000, rabbit), Diet intervention. C57BL/6 WT mice and congenic Vsig4 littermates were pro-IL-1β (#12242, CST, 1:1000, mouse), IFN-γ (#sc-52557, Santa Cruz, 1:200, rat), received NCD or HFD (#MD12031, Medicience Ltd., Nanjing, China) starting at and IL-6 (#sc-130326, Santa Cruz, 1:200, mouse) in macrophages or liver tissues as the age of 8 weeks. The body mass was evaluated every week for a total of 10 weeks. well as VAT was measured by western blot. Uncropped western blot images are The distribution of fat tissues in obese mice was made using an in vivo micro- shown in Supplementary Fig. 12. computed tomography scanner (μCT, Quantum FX, Perkin Elmer, Hopkinton, MA, USA). At the end of the feeding experiment, the mice were sacrificed, and blood was collected in EDTA-coated tubes and centrifuged to collect plasma. The Flow cytometry. To measure the mROS superoxide, macrophages were incubated liver and epididymal WAT were dissected, weighed, and immediately frozen in with MitoSOX red (5 μM, Life technologies, Eugene, Oregon, USA) at 1.0 μM for 1 liquid nitrogen. The morphometry of individual fat cells was assessed using digital h in phenol red-free DMEM (Invitrogen). The death cells were excluded firstly by image analysis as described previously . Briefly, microscopic images were digitized staining with LIVE/DEATHFixable Near-IR Ded Cell Stain Kit (Life Technologies). in 24-bit RGB (specimen level pixel size 1.28 × 1.28 μm ). Recognition of fat cells To measure the expression of activation markers on cell surface, suspended cells was initially performed by applying a region growing algorithm on manually were incubated for 1 h at room temperature in dark using fluorescent antibodies indicated seed points, and the minimum Feret diameter was calculated. (anti-B7-H1, anti-B7-DC, anti-B7-H3, and anti-CD40). To detect intracellular proinflammatory cytokines (pro-IL-1β, TNF, and IL-6) expression, macrophages or GTT and ITT experiments. For GTT, the animals were i.p. injected with 2 g/kg T cells were isolated and treated with brefeldin A for 4 h. mAbs were then added, and further incubated for an additional 1 h. all of these fluorescent antibodies were glucose (#G6125, Sigma-Aldrich, St. Louis, MO, USA) after 12 h of fasting, and blood was drawn to measure blood glucose 0, 15, 30, 60, 90 and 120 min after purchased from eBioscience (San Diego, CA, USA). A total of 10,000 live cells were analyzed by FACsAria cytometer (BD, Franklin Lakes, NJ, USA). All the flow injection. For ITT, 0.5 U/kg of insulin (Novolin R, Novo Nordisk) was i.p. injected after 6 h of fasting, and blood was drawn at 0, 15, 30, 45, and 60 min thereafter. cytometry data were analyzed using CellQuest Pro software. Seahorse XFp metabolic flux analysis. The OCR was measured using an XFp Quantitative RT-PCR. Total RNA was extracted from cultured cells or the indi- extracellular analyzer (Agilent Technologies, Santa Clara, CA, USA). Macrophages cated tissues with TRIzol reagent according to the manufacturer’s instructions were seeded at 2.0 × 10 cells/well density in 8-well plates for 5 h to allow adherence (Invitrogen). First-strand cDNA was synthesized with the PrimeScript RT-PCR Kit to the plate. After 2 h of LPS (800 ng/ml) administration, the cells were changed to (Takara, Dalian, China). The expression of mRNA encoding for the indicated genes unbuffered assay media (base medium supplemented with 10 mM glucose, 1 mM was quantified by quantitative (q)RT-PCR with the SYBR Premix ExTaq kit 12 NATURE COMMUNICATIONS 8: 1322 DOI: 10.1038/s41467-017-01327-4 www.nature.com/naturecommunications | | | NATURE COMMUNICATIONS | DOI: 10.1038/s41467-017-01327-4 ARTICLE (Takara) and was normalized to the expression of β-actin. qRT-PCR was per- p < 0.05 was considered a statistically significant difference. All results shown are formed with specific primers (Supplementary Table 2). The results were compared representative of at least three separate experiments. −ΔΔCt by the 2 method. Data availability. The authors declare that all data supporting the findings of this study are available within the article and its Supplementary Information Files or Lentiviral constructs and transduction. The mouseVsig4 (NM_177789) cDNA from the corresponding author upon request. ORF clone (#MR203780) and Pdk2 (NM_133667) cDNA ORF clone (#MG206400) were purchased from OriGene Technologies, Inc. (Rockville, MD, USA). The whole gene expression cDNA for Vsig4, Pdk2, the truncation mutants of Vsig4, and Vsig4 Received: 7 September 2016 Accepted: 11 September 2017 site-directed mutagenesis were further amplified with specific primers (Supple- mentary Table 3). cDNA was cloned into the pCDH-MCS-T2A-copGFP-MSCV (CD523A-1) vector. This vector was mutated using the QuickChange site-directed mutagenesis kit II (Stratagene, Santa Clara, CA, USA). The lentiviral packaging vectors-psPAX2 and pVSVG were purchased from Addgene (Cambridge, MA, USA). The psPAX2 plasmids (2 μg), the expression vectors (2 μg) and the pVSVG References plasmids (2 μg) were cotransfected into 293T cells, and the virus supernatants were 1. Wynn, T. A. & Vannella, K. M. Macrophages in tissue repair, regeneration, and collected after 48 h (2,000 rpm/min, 3 min). RAW264.7 cells were transduced with unconcentrated virus supernatant overnight in the presence of 8 mg/ml polybrene fibrosis. Immunity 44, 450–462 (2016). and selected in puromycin (0.5 mg/ml). The expression of VSIG4 and PDK2 was 2. Murray, P. J. & Wynn, T. A. Protective and pathogenic functions of measured by western blot. macrophage subsets. Nat. Rev. Immunol. 11, 723–737 (2011). 3. Sica, A. & Mantovani, A. Macrophage plasticity and polarization: in vivo veritas. J. Clin. Investig. 122, 787–795 (2012). Pdk2 and Stat3 shRNA silencing. Lentiviral constructs with shRNAs directed 4. Murray, P. J. et al. Macrophage activation and polarization: nomenclature and against mouse Pdk2 and Stat3 in the pGV112 vectors, which were prepared by experimental guidelines. Immunity 41,14–20 (2014). GeneChem (Shanghai, China). Primers are presented in Supplementary Table 4. 5. Olefsky, J. M. & Glass, C. K. Macrophages, inflammation, and insulin Lentivirus was prepared by transient transfection of 293T cells with transfer vectors resistance. Annu. Rev. Physiol. 72, 219–246 (2010). along with third-generation packaging constructs (pHelper 1.0 and pHelper 2.0). 6. Weisberg, S. P. et al. Obesity is associated with macrophage accumulation in The viral titers were determined with serial dilution of virus-containing media on adipose tissue. J. Clin. Investig. 112, 1796–1808 (2003). NIH3T3 cells. RAW264.7 cells were transfected with unconcentrated virus 7. Pope, M. et al. Pattern of disease after murine hepatitis virus strain 3 infection supernatant overnight in the presence of 8 mg/ml polybrene and selected in pur- correlates with macrophage activation and not viral replication. J. Virol. 69, omycin (0.5 mg/ml). 5252–5260 (1995). 8. Zhang, L. & Wang, C. C. Inflammatory response of macrophages in infection. Hepatobiliary Pancreat. Dis. Int. 13, 138–152 (2014). Chromatin immunoprecipitation and qPCR. BMDMs underwent cross-linking for 10 min with 1% formaldehyde in medium. Chromatin fragments were pre- 9. Ganeshan, K. & Chawla, A. Metabolic regulation of immune responses. Annu. pared, followed by immunoprecipitation with anti-Dnmt3a (#ab2850, Abcam, Rev. Immunol. 32, 609–634 (2014). 1:1000, Rabbit), anti-STAT3 (#ab119352, Abcam, 1:1000, mouse) or rabbit/mosue 10. Jha, A. K. et al. Network integration of parallel metabolic and transcriptional IgG isotype mAbs and coupled to Dynabeads Protein G. DNA was phenol/ data reveals metabolic modules that regulate macrophage polarization. chloroform extracted and ethanol precipitated in the presence of glycogen. DNA Immunity 42, 419–430 (2015). was purified with a PCR purification kit (#28104, Qiagen). qPCR was performed 11. Yeung, F. et al. Modulation of NF-kappaB-dependent transcription and cell with special primers (Supplementary Table 5) flanking the putative Dnmt3a- and survival by the SIRT1 deacetylase. EMBO J. 23, 2369–2380 (2004). STAT3-binding sites. The input DNA was an aliquot of sheared chromatin before 12. Kauppinen, A., Suuronen, T., Ojala, J., Kaarniranta, K. & Salminen, A. immunoprecipitation and was used for normalization of the samples to the amount Antagonistic crosstalk between NF-kappaB and SIRT1 in the regulation of chromatin added to each ChIP. of inflammation and metabolic disorders. Cell Signal. 25, 1939–1948 (2013). 13. Tannahill, G. M. et al. Succinate is an inflammatory signal that induces IL-1beta Vsig4 gene promoter methylation analyses. The Sequenom MassARRAY plat- through HIF-1alpha. Nature 496, 238–242 (2013). form (CapitalBio, Beijing, China) was used to perform the quantitative methylation 14. West, A. P. et al. TLR signalling augments macrophage bactericidal activity analysis of Vsig4. This system uses matrix-assisted laser desorption/ionization through mitochondrial ROS. Nature 472, 476–480 (2011). time-of-flight (MALDI-TOF) mass spectrometry in combination with RNA base- 15. Zhou, R., Yazdi, A. S., Menu, P. & Tschopp, J. A role for mitochondria in specific cleavage (MassCLEAVE). A detectable pattern is then analyzed for its NLRP3 inflammasome activation. Nature 469, 221–225 (2011). methylation status. PCR primers were designed with Methprimer (http:// 16. Wen, H. et al. Fatty acid-induced NLRP3-ASC inflammasome activation epidesigner.com) and were showed in (Supplementary Table 6). The spectra interferes with insulin signaling. Nat. Immunol. 12, 408–415 (2011). methylation ratios were generated by Epityper software version 1.0 (Sequenom, San 17. Michiels, C., Minet, E., Mottet, D. & Raes, M. Regulation of gene expression by Diego, CA). oxygen: NF-kappaB and HIF-1, two extremes. Free Radic. Biol. Med. 33, 1231–1242 (2002). Vsig4 gene promoter luciferase reporter assay. A840bp (-840/+1) fragment of 18. Papandreou, I., Cairns, R. A., Fontana, L., Lim, A. L. & Denko, N. C. HIF-1 Vsig4 gene promoter was cloned into the pGL3-basic expression vector using mediates adaptation to hypoxia by actively downregulating mitochondrial primers described in Supplementary Table 7. The pGL3-basic and the -840/+1 oxygen consumption. Cell Metab. 3, 187–197 (2006). Vsig4 promoter constructs were fully methylated in vitro by M.SssI following the 19. Xu, J. et al. NOTCH reprograms mitochondrial metabolism for manufacturer’s instructions (New England Biolabs (NEB), Beijing, China). The proinflammatory macrophage activation. J. Clin. Investig. 125, 1579–1590 completeness of methylation was checked by measuring the extent of protection (2015). from digestion by the restriction enzymes BstUI (NEB). RAW264.7 cells were 20. Vogt, L. et al. VSIG4, a B7 family-related protein, is a negative regulator of T cultured in 12-well plates and were transfected with these plasmids by using cell activation. J. Clin. Investig. 116, 2817–2826 (2006). amaxanucleofector (4D-Nucleofector, Lonza, Allendale, NJ, USA). At 48 h after 21. Helmy, K. Y. et al. CRIg: a macrophage complement receptor required for transfection, cells were further treated with proinflammatory mediators for an phagocytosis of circulating pathogens. Cell 124, 915–927 (2006). additional 12 h. Thereafter, cells were washed with PBS and lysed in Reporter Lysis 22. Fu, W., Wojtkiewicz, G., Weissleder, R., Benoist, C. & Mathis, D. Early Buffer (Invitrogen). Luciferase reporter activities were measured in triplicate using window of diabetes determinism in NOD mice, dependent on the complement the Dual-Luciferase reporter assay system (Promega, Madison, WI, USA) receptor CRIg, identified by noninvasive imaging. Nat. Immunol. 13, 361–368 according to the manufacturer’s protocol, and quantified using the GloMax 96-well (2012). plate luminometer (Promega). The firefly luciferase to Renilla luciferase ratios were 23. Katschke, K. J. Jr et al. A novel inhibitor of the alternative pathway of determined and were defined as the relative luciferase activity. Results are shown as complement reverses inflammation and bone destruction in experimental the mean ± s.e.m. of a representative experiment performed in triplicate. To arthritis. J. Exp. Med. 204, 1319–1325 (2007). examine the transfection efficiency, RAW264.7 cells were transfected with a 24. Chen, M., Muckersie, E., Luo, C., Forrester, J. V. & Xu, H. Inhibition of the pmaxGFP control vector and transfection efficiency was assessed by in situ GFP alternative pathway of complement activation reduces inflammation in expression according to manufacturer’s protocol (Lonza). experimental autoimmune uveoretinitis. Eur. J. Immunol. 40, 2870–2881 (2010). 25. Jung, K. et al. Protective role of V-set and immunoglobulin domain-containing Statistical analysis. Survival data from in vivo experiments were analyzed by a log-rank test performed on curves generated by GraphPad Prism 4.03 Software 4 expressed on kupffer cells during immune-mediated liver injury by inducing tolerance of liver T- and natural killer T-cells. Hepatology 56, 1838–1848 (MacKiev). For all other analysis, two-tailed, unpaired Student’s t-tests with a 95% confidence interval performed on graphs generated in GraphPad Prism were used. (2012). NATURE COMMUNICATIONS 8: 1322 DOI: 10.1038/s41467-017-01327-4 www.nature.com/naturecommunications 13 | | | ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-017-01327-4 26. Stienstra, R. et al. Inflammasome is a central player in the induction of obesity 50. Li, P. P. et al. LTB4 promotes insulin resistance in obese mice by acting on and insulin resistance. Proc. Natl. Acad. Sci. USA 108, 15324–15329 (2011). macrophages, hepatocytes and myocytes. Nat. Med. 21, 239–247 (2015). 27. Uysal, K. T., Wiesbrock, S. M., Marino, M. W. & Hotamisligil, G. S. Protection 51. Chow, S. H., Deo, P. & Naderer, T. Macrophage cell death in microbial from obesity-induced insulin resistance in mice lacking TNF-alpha function. infections. Cell. Microbiol. 18, 466–474 (2016). Nature 389, 610–614 (1997). 52. Yang, C. et al. Expression of B and T lymphocyte attenuator (BTLA) in 28. Marsden, P. A. et al. The Fgl2/fibroleukin prothrombinase contributes to macrophages contributes to the fulminant hepatitis caused by murine hepatitis immunologically mediated thrombosis in experimental and human viral virus strain-3. Gut 62, 1204–1213 (2013). hepatitis. J. Clin. Investig. 112,58–66 (2003). 53. Baylin, S. B., Belinsky, S. A. & Herman, J. G. Aberrant methylation of gene 29. Guo, S. et al. TheNLRP3 Inflammasome and IL-1beta accelerate promoters in cancer---concepts, misconcepts, and promise. J. Natl Cancer. Inst. immunologically mediated pathology in experimental viral fulminant hepatitis. 92, 1460–1461 (2000). PLoS Pathog. 11, e1005155 (2015). 54. Jaenisch, R. & Bird, A. Epigenetic regulation of gene expression: how the 30. Liu, J. et al. C5aR, TNF-alpha, and FGL2 contribute to coagulation and genome integrates intrinsic and environmental signals. Nat. Genet. 33, 245–254 complement activation in virus-induced fulminant hepatitis. J. Hepatol. 62, (2003). 354–362 (2015). 55. Prokhortchouk, E. & Defossez, P. A. The cell biology of DNA methylation in 31. Zhu, L., Zhao, Q., Yang, T., Ding, W. & Zhao, Y. Cellular metabolism and mammals. Biochim. Biophys. Acta 1783, 2167–2173 (2008). macrophage functional polarization. Int. Rev. Immunol. 34,82–100 (2015). 56. He, J. Q., Wiesmann, C. & van Lookeren Campagne, M. A role of macrophage 32. Balaban, R. S., Nemoto, S. & Finkel, T. Mitochondria, oxidants, and aging. Cell complement receptor CRIg in immune clearance and inflammation. Mol. 120, 483–495 (2005). Immunol. 45, 4041–4047 (2008). 33. Patel, M. S. & Korotchkina, L. G. Regulation of mammalian pyruvate 57. Zang, X. & Allison, J. P. To be or not to be B7. J. Clin. Investig. 116, 2590–2593 dehydrogenase complex by phosphorylation: complexity of multiple (2006). phosphorylation sites and kinases. Exp. Mol. Med. 33, 191–197 (2001). 34. Singh, S. S. et al. Targeting the PI3K/Akt signaling pathway in gastric Acknowledgements carcinoma: a reality for personalized medicine? World J. Gastroenterol. 21, Dr. Bin Li (Institute Pasteur of Shanghai, Chinese Academy of Sciences) gave invaluable 12261–12273 (2015). suggestions. This work was supported by the General Program of National Natural 35. Woolfrey, K. M. & Dell’Acqua, M. L. Coordination of protein phosphorylation Science Foundation of China (NSFC, No. 81361120400, 81222023 and 81771691), and and dephosphorylation in synaptic plasticity. J. Biol. Chem. 290, 28604–28612 The National Key Research and Development Program of China (2016YFA0502204). (2015). 36. Durant, L. et al. Diverse targets of the transcription factor STAT3 contribute to T cell pathogenicity and homeostasis. Immunity 32, 605–615 (2010). Author contribution 37. Nagase, H. & Ghosh, S. Epigenetics: differential DNA methylation in Y.C. and Y.W. designed the research and analyzed the data. J.L., S.G., and X.H. per- mammalian somatic tissues. FEBS J. 275, 1617–1623 (2008). formed most experiments. Q.N. and W.Y. prepared MHV-3 virus. The manuscript was 38. Kamata, H. et al. Reactive oxygen species promote TNFalpha-induced death prepared by L.Z., and Y.C. and Z.F. conducted the plasmid mutation. B.D. and C.Y. and sustained JNK activation by inhibiting MAP kinase phosphatases. Cell 120, conducted HFD-induced obesity. 649–661 (2005). 39. Sorbara, M. T. & Girardin, S. E. Mitochondrial ROS fuel the inflammasome. Additional information Cell Res. 21, 558–560 (2011). Supplementary Information accompanies this paper at doi:10.1038/s41467-017-01327-4. 40. Bulua, A. C. et al. Mitochondrial reactive oxygen species promote production of proinflammatory cytokines and are elevated in TNFR1-associated periodic Competing interests: The authors declare no competing financial interests. syndrome (TRAPS). J. Exp. Med. 208, 519–533 (2011). 41. Jha, M. K., Lee, I. K. & Suk, K. Metabolic reprogramming by the pyruvate Reprints and permission information is available online at http://npg.nature.com/ dehydrogenase kinase-lactic acid axis: Linking metabolism and diverse reprintsandpermissions/ neuropathophysiologies. Neurosci. Biobehav. Rev. 68,1–19 (2016). 42. Peterson, R. T. & Schreiber, S. L. Kinase phosphorylation: keeping it all in the Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in family. Curr. Biol. 9, R521–R524 (1999). published maps and institutional affiliations. 43. Yoo, H. et al. Pyruvate metabolism: a therapeutic opportunity in radiation- induced skin injury. Biochem. Biophys. Res. Commun. 460, 504–510 (2015). 44. Jha, M. K. et al. Metabolic connection of inflammatory pain: pivotal role of a pyruvate dehydrogenase kinase-pyruvate dehydrogenase-lactic acid axis. J. Open Access This article is licensed under a Creative Commons Neurosci. 35, 14353–14369 (2015). Attribution 4.0 International License, which permits use, sharing, 45. Tan, Z. et al. Pyruvate dehydrogenase kinase 1 participates in macrophage adaptation, distribution and reproduction in any medium or format, as long as you give polarization via regulating glucose metabolism. J. Immunol. 194, 6082–6089 appropriate credit to the original author(s) and the source, provide a link to the Creative (2015). Commons license, and indicate if changes were made. The images or other third party 46. Vandanmagsar, B. et al. The NLRP3 inflammasome instigates obesity-induced material in this article are included in the article’s Creative Commons license, unless inflammation and insulin resistance. Nat. Med. 17, 179–188 (2011). indicated otherwise in a credit line to the material. If material is not included in the 47. Kanda, H. et al. MCP-1 contributes to macrophage infiltration into adipose article’s Creative Commons license and your intended use is not permitted by statutory tissue, insulin resistance, and hepatic steatosis in obesity. J. Clin. Investig. 116, regulation or exceeds the permitted use, you will need to obtain permission directly from 1494–1505 (2006). 48. Traves, P. G. et al. Relevance of the MEK/ERK signaling pathway in the the copyright holder. To view a copy of this license, visit http://creativecommons.org/ metabolism of activated macrophages: a metabolomic approach. J. Immunol. licenses/by/4.0/. 188, 1402–1410 (2012). 49. Wu, Y. et al. Activation of AMPK alpha 2 in adipocytes is essential for nicotine- © The Author(s) 2017 induced insulin resistance in vivo. Nat. Med. 21, 373–382 (2015). 14 NATURE COMMUNICATIONS 8: 1322 DOI: 10.1038/s41467-017-01327-4 www.nature.com/naturecommunications | | |

Journal

Nature CommunicationsSpringer Journals

Published: Nov 6, 2017

There are no references for this article.