Mice Deficient in Angiopoietin-like Protein 2 (Angptl2) Gene Show Increased Susceptibility to Bacterial Infection Due to Attenuated Macrophage Activity*

Macrophages play crucial roles in combatting infectious disease by promoting inflammation and phagocytosis. Angiopoietin-like protein 2 (ANGPTL2) is a secreted factor that induces tissue inflammation by attracting and activating macrophages to produce inflammatory cytokines in chronic inflammation-associated diseases such as obesity-associated metabolic syndrome, atherosclerosis, and rheumatoid arthritis. Here, we asked whether and how ANGPTL2 activates macrophages in the innate immune response. ANGPTL2 was predominantly expressed in proinflammatory mouse bone marrow-derived differentiated macrophages (GM-BMMs) following GM-CSF treatment relative to anti-inflammatory cells (M-BMMs) established by M-CSF treatment. Expression of the proinflammatory markers IL-1β, IL-12p35, and IL-12p40 significantly decreased in GM-BMMs from Angptl2-deficient compared with wild-type (WT) mice, suggestive of attenuated proinflammatory activity. We also report that ANGPTL2 inflammatory signaling is transduced through integrin α5β1 rather than through paired immunoglobulin-like receptor B. Interestingly, Angptl2-deficient mice were more susceptible to infection with Salmonella enterica serovar Typhimurium than were WT mice. Moreover, nitric oxide (NO) production by Angptl2-deficient GM-BMMs was significantly lower than in WT GM-BMMs. Collectively, our findings suggest that macrophage-derived ANGPTL2 promotes an innate immune response in those cells by enhancing proinflammatory activity and NO production required to fight infection.

The mammalian immune system is critical to combat infectious pathogens and comprises two branches, the innate and acquired immune systems (1). Unlike the acquired system, innate immunity is activated to provide an immediate response to pathogens (2). Thus, it constitutes the first line of host defense against pathogens and is mediated by phagocytes such as macrophages. In addition, development of inflammation is important for host resistance to infection (1), and proinflammatory macrophages are crucial to support resistance to intracellular bacteria and control the acute phase of infection (3)(4)(5)(6)(7)(8)(9)(10).
We reported previously that ANGPTL2 4 signaling functions in tissue repair through adaptive inflammation and angiogenesis (11,12), whereas excess ANGPTL2 signaling causes chronic maladaptive inflammation and subsequent pathological irreversible tissue remodeling, leading to the development of various diseases such as obesity-associated metabolic disease, atherosclerotic vascular disease, rheumatoid arthritis, and some cancers (11,(13)(14)(15). In these pathologies, ANGPTL2 activates macrophages, which participate in disease progression, suggesting that ANGPTL2 promotes proinflammatory phenotypes and stimulates inflammatory cytokine secretion by macrophages. Although these activities are associated with disease, inflammation is required for the innate immune system to fight infection.
In this study, we asked whether and how ANGPTL2 functions in the innate immune response. We found that ANGPTL2 is predominantly expressed in proinflammatory bone marrowderived differentiated macrophages (GM-BMMs) established by treatment of mouse bone marrow cells with GM-CSF, and proinflammatory activity was attenuated in Angptl2-deficient relative to wild-type (WT) macrophages. In addition, we showed that ANGPTL2 activates macrophages through integrin ␣5␤1 rather than paired immunoglobulin-like receptor B (PIR-B). Interestingly, Angptl2-deficient mice were more susceptible to Salmonella enterica serovar Typhimurium infection than were WT mice. Moreover, nitric oxide (NO) production and phagocytosis ability by Angptl2-deficient GM-BMMs were significantly lower than in WT controls. These findings suggest that macrophage-derived ANGPTL2 stimulates innate immune defense capacity in macrophages by enhancing proinflammatory activity and NO production against infection.

ANGPTL2 Treatment Promotes Proinflammatory Phenotypes in Macrophages in Vitro-
To determine whether ANG-PTL2 stimulates differentiation of pro-or anti-inflammatory macrophages, we undertook two in vitro tests of differentiation (Fig. 1A). In the first, we incubated bone marrow cells from wild-type adult mice with M-CSF to induce differentiated macrophages (M-BMMs), which exhibit anti-inflammatory phenotypes (16); in the other, we incubated bone marrow cells with GM-CSF to induce proinflammatory GM-BMMs (17,18) (Fig. 1A). We then assessed expression of pro-or anti-inflammatory markers in both groups by quantitative real time PCR analysis (Fig. 1B). Relative levels of transcripts encoding proinflammatory markers such as IL-1␤, IL-12p35, IL-12p40, IL-6, and TNF-␣ were significantly increased in GM-BMMs compared with M-BMMs, whereas transcripts encoding anti-inflammatory markers such as IL-10 and the chemokine receptor CX3CR1 were significantly up-regulated in M-BMMs compared with GM-BMMs (Fig. 1B) as anticipated. Next, we examined endogenous ANGPTL2 expression in these cells and found that GM-BMMs expressed higher ANGPTL2 levels than did M-BMMs, suggesting that ANGPTL2 expression is associated with proinflammatory activity (Fig. 1C).
Next, we compared gene expression profiles in GM-BMMs from WT versus Angptl2 KO mice. Transcripts encoding proinflammatory markers such as IL-1␤, IL-12p35, and IL-12p40, but not IL-6 and TNF-␣, were significantly decreased, whereas transcripts encoding the anti-inflammatory marker CX3CR1, but not IL-10, were significantly up-regulated in GM-BMMs   from Angptl2 KO (Fig. 1D) compared with WT mice. These findings suggest that ANGPTL2 secretion by macrophages promotes proinflammatory phenotypes seen in those cells in vitro. Integrin ␣5␤1 Functions in Proinflammatory ANGPTL2 Signaling in GM-BMMs-To further assess ANGPTL2 signaling in inflammation, we conducted experiments supplementing cell cultures with recombinant mouse ANGPTL2 protein (rANGPTL2) and found that Angptl2 KO GM-BMMs treated with rANGPTL2 showed up-regulated expression of proinflammatory markers, including IL-6 and TNF-␣ ( Fig. 2A). Next, we asked what receptor might function in promoting proinflammatory phenotypes in GM-BMMs. The immune inhibitory human leukocyte immunoglobulin-like receptor B2, which contributes to hematopoietic stem cell stemness and leukemia development, reportedly functions as an ANGPTL2 receptor (19). PIR-B, its mouse orthologue (35), is reportedly expressed on and functions in immune cells (20), and Pir-b KO mice show heightened susceptibility to Salmonella infection (21). To determine whether PIR-B functioned as an ANGPTL2 receptor in promoting proinflammatory phenotypes in macrophages, we asked whether rANGPTL2 treatment could activate GM-BMMs established from Pir-b KO mice. rANGPTL2 treatment increased induction of proinflammatory transcripts such as IL-1␤, IL-12p35, IL-12p40, IL-6, and TNF-␣ in GM-BMMs from Pir-b KO mice relative to untreated Pir-b KO controls (Fig. 2B), suggesting that in this context PIR-B does not mediate ANGPTL2-dependent proinflammatory signaling.

Pir-b KO GM-BMMs
Pir-b KO GM-BMMs + ANGPTL2 Relative mRNA expression/18s Relative mRNA expression/18s FIGURE 2. Integrin ␣5␤1 partially mediates ANGPTL2-dependent proinflammatory phenotypes in macrophages. A, relative expression of mRNAs encoding proinflammatory markers in cultured GM-BMMs from Angptl2 KO mice incubated with or without rANGPTL2 (n ϭ 4). Levels in untreated cells were set at 1. B, relative expression of transcripts encoding proinflammatory markers in cultured GM-BMMs from Pir-b KO mice incubated with or without rANGPTL2 (n ϭ 4). Levels in untreated cells were set at 1. C, adhesion of GM-BMMs from WT mice to culture dishes coated with various concentrations of rANGPTL2 (n ϭ 4). D, adhesion of GM-BMMs with rANGPTL2 (20 g/ml) from WT mice preincubated with or without 25 g/ml ␣5␤1 antibody (n ϭ 4). As a negative control, adhesion was assayed in the presence of 10 mM EDTA, which inhibits integrin binding. E, relative expression of transcripts encoding proinflammatory markers in cultured GM-BMMs from Pir-b KO mice treated with rANGPTL2 with or without anti-integrin ␣5␤1 antibody (n ϭ 4). Levels in GM-BMMs not treated with antibody were set at 1. *, p Ͻ 0.05; **, p Ͻ 0.01. Data are expressed as means Ϯ S.E. (error bars) from two different experiments. body (Fig. 2D). Next, we asked whether treatment of GM-BMMs with that antibody would block proinflammatory signaling induced by ANGPTL2. We observed that treatment of GM-BMMs with integrin ␣5␤1 antibody partially inhibited induction of transcripts encoding the proinflammatory markers IL-1␤, IL-12p35, IL-12p40, IL-6, and TNF-␣ by ANGPTL2 (Fig. 2E), suggesting that integrin ␣5␤1 functions in the macrophage response to ANGPTL2.
Angptl2 KO Mice Are More Susceptible to Salmonella Infection than Are WT Mice-Proinflammatory macrophages mediate resistance to intracellular pathogens (22)(23)(24). To evaluate ANGPTL2 function in proinflammatory macrophages, we generated in vivo models of WT or Angptl2 KO mice infected with the Salmonella enterica serovar Typhimurium LT2, a Gramnegative, facultative intracellular pathogen that grows outside and inside of host cells. Mutant and WT mice were intraperitoneally infected with various LT2 doses (from 2 ϫ 10 4 to 2 ϫ 10 5 colony-forming units (cfu)/mouse) in groups of 10 per dose, and their survival was monitored for 4 weeks thereafter. In both groups, mice inoculated with high doses (2 ϫ 10 5 cfu/mouse) died within 9 days of infection ( Fig. 3A, left panel). At middle doses (7 ϫ 10 4 cfu/mouse; Fig. 3A, middle panel), differences in survival were more apparent: many infected Angptl2 KO mice showed signs of morbidity sooner than did WT mice during the course of infection, and all Angptl2 KO mice died by 6 days after infection. By contrast, only half of the WT mice died during the same period (Fig. 3A, middle panel). Plasma ANGPTL2 levels in LT2-infected WT mice were significantly increased by day 3 of infection ( Fig. 3B). At low doses (2 ϫ 10 4 cfu/mouse), all mice of either genotype survived except one Angptl2 KO mouse (Fig.  3A, right panel). Next, to determine tissue localization of inoculated bacteria, we sacrificed mice receiving the midlevel dose (7 ϫ 10 4 cfu/mouse) 3 days after infection and analyzed bacterial loads in organs such as blood, liver, and spleen. Angptl2 KO mice showed higher bacterial loads in liver than did WT mice (Fig. 3C), suggesting that lowered immune defense seen in Angptl2 KO mice may be associated with impaired bacterial replication.
Angptl2 KO GM-BMMs Show Impaired NO Production-Next, we measured NO production by GM-BMMs, an activity required for antibacterial responses (25)(26)(27)(28). In these experiments, we estimated NO production by assaying for nitrite, a stable metabolite of NO (29,30). GM-BMMs from Angptl2 KO mice showed less nitrite production than did WT GM-BMMs at 6 and 12 h following exposure to LT2 at a multiplicity of infection (m.o.i.) of 1 (Fig. 4, A and B), supporting the idea that ANGPTL2 modulates bacterial clearance capacity through NO production. However, based on nitrite analysis, the difference in NO production by GM-BMMs in WT versus Angptl2 KO mice was lesser at 12 h than at 6 h (Fig. 4, A  and B), suggesting that bacterial clearance by macrophages is more impaired at early phases of infection in Angptl2 KO relative to WT mice.    5A). However, phagocytotic activity of GM-BMMs from Angptl2 KO mice decreased relative to that in WT mice at higher m.o.i. values (Ͼ10) (Fig. 5A), suggesting that ANGPTL2 functions to control Salmonella infection by modulating phagocytosis. Therefore, we next examined expression of receptors that function in phagocytosis, namely the toll-like receptor 4 (TLR4) (CD284), Fc-␥ receptor (CD16/32), scavenger receptor (CD36), and complement receptor (CD21/35), in GM-BMMs from WT or Angptl2 KO mice. We observed no differences in expression of these receptors in GM-BMMs from mice of either genotype (Fig. 5B). Furthermore, when we assayed phagocytotic activity of GM-BMMs from WT or Angptl2 KO mice using fluorescent beads, we observed no differences in GM-BMMs from WT or Angptl2 KO mice (Fig. 5C). We conclude that potential changes in expression of major receptors associated with phagocytosis cannot account for impaired bacterial clearance seen in Angptl2 KO GM-BMMs.
Angptl2 KO Peritoneal Macrophages Show Impaired Migration and NO Production-The studies above address activities in GM-BMMs derived from WT or Angptl2 KO mice. Next, we assessed responses of primary peritoneal macrophages derived from WT and Angptl2 KO mice. In unstimulated conditions, the number of peritoneal macrophages was comparable in mice of both genotypes (Fig. 6A). However, following thioglycollate stimulation, the number of peritoneal macrophages was significantly lower in Angptl2 KO relative to WT mice (Fig. 6B). We then evaluated responses of peritoneal macrophages to in vitro LT2 infection. As observed in GM-BMMs (Fig. 5), we observed comparable phagocytosis in peritoneal macrophages of both genotypes at an m.o.i. of 1 (Fig. 6C). We also examined nitrite production in response to LT2 infection in peritoneal macrophages derived from mice of both genotypes and found that nitrite production by peritoneal macrophages from Angptl2 KO mice was decreased relative to that seen in WT mice similar to our observations in GM-BMMs (Fig. 6D). Next, we compared gene expression profiles in peritoneal macrophages following lipopolysaccharide (LPS) treatment. Expression of the proinflammatory markers IL-1␤ and IL-12p40 was significantly lower in peritoneal macrophages from Angptl2 KO mice (Fig.  6E). We also observed relatively lower TLR4 expression in peritoneal macrophages from Angptl2 KO mice (Fig. 6F), suggesting that decreased TLR4 expression may underlie attenuation of macrophage activity seen in Angptl2 KO mice.

Discussion
ANGPTL2 is a proinflammatory mediator that induces inflammation-associated diseases such as obesity-associated metabolic disease, atherosclerotic vascular disease, and rheumatoid arthritis (11,(13)(14)(15). Its role in infection, however, had not previously been addressed. Here, we focused on the function of macrophage-derived ANGPTL2 in infectious disease as macrophages are key inflammatory cells in this context. We found that ANGPTL2 was predominantly expressed in macrophages with proinflammatory phenotypes, and macrophagederived ANGPTL2 likely activated other macrophages to acquire an inflammatory phenotype. Angptl2 KO mice were also more susceptible to infection.
Our analysis shows that GM-BMMs, which exhibit proinflammatory macrophage phenotypes, express higher levels of ANGPTL2 than do anti-inflammatory M-BMMs. We also found that exogenous rANGPTL2 can rescue expression of mRNAs encoding inflammatory cytokines, including IL-1␤, IL-12p35, IL-12p40, IL-6, and TNF-␣, in Angptl2 KO GM-BMMs, suggesting that the macrophage-derived ANGPTL2 mediates proinflammatory phenotypes in an autocrine manner. Conversely, levels of mRNAs encoding IL-6 and TNF-␣ were comparable in Angptl2 KO and WT GM-BMMs in the absence of exogenous rANGPTL2. Thus, other mechanisms may regulate IL-6 and TNF-␣ expression in Angptl2 KO GM-BMMs.
We reported previously that integrin ␣5␤1 is an ANGPTL2 receptor and functions in inflammatory signaling (31), whereas others have reported that PIR-B, the mouse orthologue of human leukocyte immunoglobulin-like receptor B2, also acts as an ANGPTL2 receptor (19,32). Pir-b KO mice are susceptible to Salmonella infection (21), raising the possibility that Angptl2 signaling through PIR-B mediates this function. However, here we found that supplementation of recombinant ANGPTL2 protein activated expression of proinflammatory cytokines in GM-BMMs from Pir-b KO mice, an effect blocked in part by treatment with anti-integrin ␣5␤1 antibodies. These findings show that ANGPTL2 mediates proinflammatory phenotypes in macrophages in part through the integrin ␣5␤1 rather than the PIR-B receptor. One possibility is that an ANGPTL2/PIR-B pathway might serve a different function in GM-BMMs. In contrast, PIR-B is reportedly an inhibitory receptor, whereas the homologous PIR-A receptor is thought to be activating (33). Thus, ANGPTL2 might bind PIR-A. However, Zhang et al. (19) reported that ANGPTL2 does not bind to leukocyte immunoglobulin-like receptor A, an orthologue of mouse PIR-A. Fur-

At 6h after infection At 12h after infection
Nitrite production (μM) Nitrite production (μM) ther studies are needed to clarify ANGPTL2/PIR-A or -PIR-B signaling in this context. We also show here that expression of receptors associated with phagocytosis such as TLR4, Fc-␥ receptor, scavenger receptor, and complement receptor was comparable in GM-BMMs derived from WT and Angptl2 KO mice. Conversely, Angptl2 KO GM-BMMs show impaired phagocytosis of bacteria at higher m.o.i. values (Ͼ10). Phagocytosis is triggered by pattern recognition receptors and is accompanied by actin polymerization and activation of the GTPases Rac1, Rac2, and Cdc42 to enable formation of the phagocytic cup (34). We reported previously that ANGPTL2 activates the integrin ␣5␤1/Rac pathway (11). Thus, ANGPTL2 may promote phagocytosis by activating actin polymerization through that pathway.
In addition to phagocytosis, activated macrophages also function to combat infection caused by intracellular pathogens. NO is a critical effector of macrophages in antipathogenic responses, and NO harbors direct antimicrobial activity (35)(36)(37). Here, using nitrite levels as a readout, we found that NO production by Angptl2-deficient macrophages in response to LT2 infection was lower than that seen in WT macrophages. We conducted a preliminary analysis to adoptively transfer WT macrophages into the peritoneum of LT2-infected Angptl2 KO mice and observed that transfer tended to rescue survival (data not shown). These studies suggest that intracellular bacterial clearance, rather than phagocytosis, is impaired in Angptl2 KO macrophages, but these conclusions require confirmation in future studies.
IL-1␤ family factors such as IL-1␤ and IL-33 induce NO in macrophages (38). Given that macrophages from Angptl2 KO mice show attenuated expression of proinflammatory cytokines, including IL-1␤, we conclude that this decrease underlies reduced NO production seen in macrophages from mutant mice. In addition, IFN-␥ activates macrophages and promotes expression of antimicrobial effectors, including NO (39).
Proinflammatory polarization of macrophages requires Th1 cytokines, particularly T cell-derived IFN-␥ (40). At day 3 after LT2 infection in vivo, we observed lower levels of serum IFN-␥ in Angptl2 KO compared with WT mice (data not shown). As our primary focus here is on macrophage function, we did not evaluate T cells in detail for this study. However, levels of IFN-␥ produced by T cells may be reduced in Angptl2 KO relative to WT mice, a possibility to be addressed in future studies.
In summary, we report that ANGPTL2 is predominantly expressed in proinflammatory macrophages and that macrophage-derived ANGPTL2 promotes proinflammatory phenotypes in part through the integrin ␣5␤1 receptor. Angptl2 KO mice are likely susceptible to infection due to attenuated NO production by Angptl2-deficient macrophages. These findings suggest an important role for ANGPTL2 in NO production by macrophages; however, further studies are necessary to deter- mine whether ANGPTL2 directly increases NO production in these cells.

Experimental Procedures
Mice-Female Angptl2-deficient (Angptl2 KO) and WT littermate mice on a C57BL/6N background were used for all experiments as described (11). Pir-b-deficient (Pir-b KO) mice were obtained from Dr. Takai (41). All procedures were approved by the Kumamoto University Ethics Review Committee for Animal Experimentation.
Culture of Bone Marrow Cells-Bone marrow-derived macrophages grown in M-CSF (eBioscience, San Diego, CA) or GM-CSF (PeproTech, Rocky Hill, NJ) were generated as described (42). Briefly, bone marrow cells were obtained from mouse femurs. After lysing erythrocytes, cells were cultured at 1 ϫ 10 6 cells/ml in RPMI 1640 medium containing 10% FCS and 50 ng/ml M-CSF or 20 ng/ml GM-CSF conditioned medium for 2 days. After removing nonadherent cells, adherent cells were cultured in fresh medium for 4 more days and harvested for use as M-BMMs in the case of M-CSF-treated cells or GM-BMMs in the case of GM-CSF-treated cells.
Collection of Peritoneal Macrophages-After euthanizing mice, the outer skin of the peritoneum was cut, 5 ml of ice-cold phosphate-buffered saline (PBS) (with 3% FCS) was injected into the peritoneal cavity, the peritoneum was gently massaged, and the perfusate was collected. The perfusate was then centrifuged at 2000 rpm for 5 min, and the supernatant was discarded. Precipitated cells were resuspended in RPMI 1640 medium containing 10% FCS. For induction analysis, mice were administered 4% thioglycollate (BD Biosciences) intraperitoneally. Resultant peritoneal exudate cells were collected at 4 days after treatment by 4% thioglycollate (43).
rANGPTL2 Treatment-Recombinant mouse ANGPTL2 hexahistidine-tagged protein was expressed in Escherichia coli Rosetta TM pLacI cells (Merck) in inclusion bodies. Briefly, the bacterial cultures were grown in LB medium at 37°C to midlog phase, and the protein was induced with isopropyl 1-thio-␤-Dgalactopyranoside at 13°C for 18 h. The harvested cells (6 g) were suspended in 30 ml of ice-cold buffer (20 mM Hepes-NaOH, pH 7.4, 0.5 M NaCl, 0.05% Triton X-100, 1 mM PMSF) and thoroughly disrupted by sonication. The inclusion bodies were solubilized, reduced, and modified by 3-trimethylammoniopropyl methanethiosulfonate bromide (Wako, Osaka, Japan) according to a published procedure (44) with minor modifications. 3-Trimethylammoniopropyl methanethiosulfonate bromide-modified proteins were desalted on a Sephadex G-25 column (GE Healthcare). Desalted proteins were loaded onto a Talon column (Takara) and eluted with 0.15 M imidazole after washing the column with solubilizing buffer. The eluted sample was desalted again, and a portion of the sample was diluted into refolding buffer (2 g/ml final concentration) containing 2 mM cysteine and 0.5 mM cystine at 4°C for 14 h. Proteins were adsorbed onto a Source 30 reverse-phase matrix (GE Healthcare) and eluted with acetonitrile containing 0.04% trifluoroacetic acid. The eluate was freeze-dried, dissolved in 0.1% acetic acid, and stored at Ϫ80°C. The samples had an endo-toxin level less than 0.3 endotoxin unit/mg. GM-BMMs from Angptl2 or Pir-b KO mice were plated at 2 ϫ 10 5 cells/well in RPMI 1640 medium containing 10% FCS plus 20 ng/ml GM-CSF conditioned medium. Medium was then changed to fresh medium containing 1 g/ml rANGPTL2 protein, and cells were then cultured for more than 6 h before analysis of inflammatory cytokine gene expression by quantitative real time PCR.
Integrin Assay-GM-BMMs from Angptl2 KO mice were plated at 2 ϫ 10 5 cells/well in RPMI 1640 medium containing 10% FCS plus 20 ng/ml GM-CSF conditioned medium with or without anti-mouse integrin ␣5␤1 antibody (Merck Millipore, Darmstadt, Germany) for 30 min. Medium was then changed to fresh medium containing 1 g/ml rANGPTL2 protein, and cells were then cultured for 6 h before analysis of inflammatory cytokine gene expression by quantitative real time PCR.
Cell Adhesion Assay-Various concentrations of ANGPTL2 protein were coated onto 96-well flat bottomed plates overnight at 4°C and blocked with 3% BSA for 1 h at 37°C. GM-BMMs from WT mice were adjusted to 1 ϫ 10 5 cells/ml and preincubated in serum-free medium with or without anti-␣5␤1 (Merck Millipore) for 30 min at 37°C. Pretreated cells were incubated in the ANGPTL2 protein-coated dishes at 37°C for 1 h. Non-adherent cells were removed by gentle PBS washing, and then adherent cells were fixed with 4% paraformaldehyde in PBS for 30 min and stained with 0.5% crystal violet in 25% methanol for 30 min. Plates were rinsed with tap water, stained cells were solubilized in 1% SDS, and the A 595 value was determined.
Salmonella Infection-Mice were administered Salmonella enterica serovar Typhimurium LT2 in 0.1 ml of PBS intraperitoneally (i.p.). At various time points thereafter, body weights and mouse survival were monitored, and mice were sacrificed to obtain liver, spleen, and blood samples to assess bacterial growth. Organs were drawn into a heparin-coated tube and weighed. Liver and spleen were homogenized in 0.1% deoxycholic acid in PBS, and homogenates were serially diluted and subjected to a cfu assay on LB agar plates.
Quantitation of ANGPTL2 Protein-Plasma ANGPTL2 protein was quantified by ELISA. ANGPTL2 concentration was estimated using an ANGPTL2 Assay kit (Immuno-Biological Laboratories, Gunma, Japan) according to the manufacturer's instructions.
Nitrite Production Assay-GM-BMMs or peritoneal macrophages (1 ϫ 10 5 cells) were plated onto 96-well plates in DMEM containing 10% FCS and stimulated with LT2 at an m.o.i. of 1. Cells were then incubated at 37°C for various time periods under 5% CO 2 before supernatants were collected. Accumulation of nitrite, a stable metabolite of NO, in supernatants was quantified by a Griess reaction as described (29,30). Briefly, 50-l aliquots of culture supernatants were dispensed in triplicate into 96-well plates and mixed with 25 l of Griess reagent A (1% sulfanilamide in 5% H 3 PO 4 ). After incubation for 5 min, 25 l of Griess reagent B (0.1% N-(1-naphthyl)-ethylenediamine) was added and incubated for 10 min at room temperature. Sample absorbance at 540 nm was compared with that of a sodium nitrite standard using a microplate reader (Bio-Rad).
Bacterial Phagocytosis Assay-For in vitro infection, LT2 (see above) was added at the indicated m.o.i. into 96-well plates containing GM-BMMs or peritoneal macrophages (1 ϫ 10 5 cells/well). Cultures were centrifuged briefly and then incubated at 37°C for 20 min under 5% CO 2 before addition of gentamicin (Wako) at a final concentration of 100 g/ml for 30 min to kill extracellular LT2. After three PBS washes, infected cells were lysed in 0.1% deoxycholic acid in PBS before cfu plate counts were undertaken (45).
Bead Phagocytosis Assay-GM-BMMs from WT or Angptl2 KO mice were plated onto 24-well plates at 5 ϫ 10 5 cells/well in RPMI 1640 medium containing 10% FCS and incubated overnight at 37°C. The next day, bead phagocytosis was quantified using a phagocytosis assay kit (Cayman, Ann Arbor, MI) according to the manufacturer's instructions. All samples were analyzed by BD Accuri C6 flow cytometry after samples were treated with assay buffer and vortexed (46).
Peritoneal Macrophage Cell Counts-Peritoneal macrophages from WT or Angptl2 KO mice were pretreated with Fc blocker (Bio Legend) and incubated with APC-anti-F4/80 (Bio Legend). After incubation, counting beads (CountBright absolute counting beads, Invitrogen) were added as a reference to calculate absolute cell numbers. Cells were analyzed by BD Accuri C6 flow cytometry (47).
LPS Stimulation-Peritoneal macrophages from WT or Ang-ptl2 KO mice were plated at 2 ϫ 10 5 cells/well in RPMI 1640 medium containing 10% FCS. Cells were cultured with or without 10 g/ml LPS (Sigma) for 6 h before analysis of gene expression by quantitative real time PCR.
Statistical Analysis-Data were recorded as the mean Ϯ S.E. Differences in group survival were analyzed using Mantel-Cox log rank p test. All other simple comparisons were performed using Student's t test with p Ͻ 0.05 considered statistically significant.