A Myeloid Hypoxia-inducible Factor 1α-Krüppel-like Factor 2 Pathway Regulates Gram-positive Endotoxin-mediated Sepsis*

Background: Gram-positive bacterial infections and sepsis are a significant cause of morbidity and mortality. Results: KLF2 inhibits Gram-positive, bacterial endotoxin-induced HIF-1α expression and macrophage activation. Conclusion: Transcription factors KLF2 and HIF-1α are critical regulator of Gram-positive sepsis. Significance: Pharmacological agents that modulate the KLF2/HIF-1 pathway may allow for therapeutic gain in the treatment of bacterial infections and sepsis. Although Gram-positive infections account for the majority of cases of sepsis, the molecular mechanisms underlying their effects remains poorly understood. We investigated how cell wall components of Gram-positive bacteria contribute to the development of sepsis. Experimental observations derived from cultured primary macrophages and the cell line indicate that Gram-positive bacterial endotoxins induce hypoxia-inducible factor 1α (HIF-1α) mRNA and protein expression. Inoculation of live or heat-inactivated Gram-positive bacteria with macrophages induced HIF-1 transcriptional activity in macrophages. Concordant with these results, myeloid deficiency of HIF-1α attenuated Gram-positive bacterial endotoxin-induced cellular motility and proinflammatory gene expression in macrophages. Conversely, Gram-positive bacteria and their endotoxins reduced expression of the myeloid anti-inflammatory transcription factor Krüppel-like transcription factor 2 (KLF2). Sustained expression of KLF2 reduced and deficiency of KLF2 enhanced Gram-positive endotoxins induced HIF-1α mRNA and protein expression in macrophages. More importantly, KLF2 attenuated Gram-positive endotoxins induced cellular motility and proinflammatory gene expression in myeloid cells. Consistent with these results, mice deficient in myeloid HIF-1α were protected from Gram-positive endotoxin-induced sepsis mortality and clinical symptomatology. By contrast, myeloid KLF2-deficient mice were susceptible to Gram-positive sepsis induced mortality and clinical symptoms. Collectively, these observations identify HIF-1α and KLF2 as critical regulators of Gram-positive endotoxin-mediated sepsis.

iNOS in various organs, significantly elevated plasma levels of TNF-␣, and caused multiorgan dysfunction syndrome.
Recent mechanistic efforts have also begun to shed light on the cellular events that mediate the effects of Gram-positive bacterial wall products. For example, studies from Dunne et al. (13) demonstrated that the type 1 macrophage scavenger receptor binds to LTA through polyanionic bonds. Studies by Schwandner et al. (14) indicated that LTA induced cellular activation via Toll-like receptors (e.g. TLR2) and NF-B activation. Consistent with these observations, in vitro and in vivo studies indicated that Gram-positive bacterial products can induce expression of proinflammatory cytokines (e.g. TNF-␣, INF-␥, IL-1␤, IL-6) and iNOS in myeloid cells that are key contributors to the sepsis syndrome. In keeping with these observations, studies using a human whole blood model indicated that lipoteichoic acid can induce expression of TNF-␣, IL-6, and IL-10 in a time-and dose-dependent manner (15). Collectively, these studies indicate that precise molecular mechanisms mediate the effects of Gram-positive bacterial endotoxins that contribute to the systemic inflammatory response syndrome.
In addition to bacterial wall components, sites of infection are typically characterized by hypoxia. The importance of hypoxia has been highlighted by studies focused on the role of the master regulator of hypoxic signaling, hypoxia-inducible factor 1␣ (HIF-1␣). Studies by Johnson and co-workers (16 -18), largely through loss of function approaches in vitro and in vivo, have revealed that myeloid HIF-1␣ augments proinflammatory cytokine expression and alters macrophage metabolic activity and bacterial killing. HIF-1 is a heterodimeric helixloop-helix transcription factor whose expression is stringently regulated at mRNA and protein levels. Protein stability of the ␣ subunit of HIF-1 is regulated by a family of enzymes termed prolylhydroxylases, whose action directs HIF-1␣ degradation by the ubiquitin-proteasome pathway in a process dependent on interaction with von Hippel-Lindau (VHL) tumor suppressor protein (19). Under hypoxic conditions, prolylhydroxylase activity is inhibited, and HIF-1␣ accumulates and translocates into the nucleus, where it binds the constitutively activated HIF-1␤. The resultant heterodimer HIF-1 binds to the hypoxic response elements (HREs) of target genes (20). In addition to hypoxia, lipopolysaccharide can potently induce HIF-1␣ transcription. Thus, the combination of bacterial products and hypoxia can, through both transcriptional and posttranscriptional mechanisms, augment HIF-1␣ levels (21).
Krüppel-like factors (KLFs) are a subclass of the zinc finger family of DNA-binding transcription factors implicated in a broad spectrum of biological processes. Krüppel-like factor 2 (KLF2) was initially identified by the Lingrel laboratory and termed lung Krüppel-like factor because of its high level in lung tissues (22). In the context of myeloid cell biology, our group identified KLF2 as a negative regulator of myeloid cell activation (23). KLF2 overexpression was found to attenuate LPSinduced expression of cytokines and chemokines through inhibiting transcriptional activity of NF-B and activator protein 1 (AP-1). More recently we found that myeloid KLF2 regulates the host response to polymicrobial infection and lipopolysaccharide-induced septic shock (24). However, a role for this factor in Gram-positive infection and sepsis has not been elu-cidated. Herein, we identify HIF-1␣ as an important mediator of Gram-positive bacterial endotoxin-induced cellular functions and inflammatory gene expression. Further, we provide evidence that KLF2 modulate Gram-positive endotoxin-induced HIF-1␣ expression and myeloid cell response in vitro and in vivo.

EXPERIMENTAL PROCEDURES
Materials-Lipopolysaccharide was purchased from Sigma-Aldrich (St. Louis, MO). Teichoic acid was obtained from Gen-Way Biotech, Inc. (San Diego, CA). Anti-HIF-1␣ antibody and anti-actin antibody were obtained from Novus Biologicals (Littleton, CO) and Santa Cruz Biotechnology, Inc. (Santa Cruz, CA), respectively. Transwell permeable supports with 8.0-m pore size were obtained from Corning, Inc. (Lowell, MA). Escherichia coli (ATCC number 21149) and Staphylococcus aureus (strain Newman) bacterial strains were obtained from the ATCC. Lipoteichoic acid was obtained from Sigma-Aldrich or extracted as described previously (25). Ad-GFP (control) and Ad-KLF2 (KLF2) adenoviral constructs were generated by the Harvard Gene Therapy Group as described previously (26). All other chemicals and reagents used were of analytical grade and were obtained from commercial sources.
Cell Culture-The RAW264.7 cell line was purchased from the ATCC and cultured in DMEM supplemented with 10% FBS, 100 units/ml penicillin, 100 g/ml streptomycin, and 2 mM glutamine in a humidified atmosphere of 5% CO 2 and 95% air at 37 ºC. Primary mouse macrophages and neutrophils were obtained from the peritoneal cavity by inducing peritonitis with 3% thioglycolate broth in 8-to 12-week-old mice as described previously (16). These primary peritoneal macrophages and neutrophils were cultured in serum supplemented DMEM as described above.
Experimental Mouse Models-The mouse lines used in this study were generated as described previously (24). Briefly, a mouse line expressing lysozyme M promoter-driven Cre recombinase (LysM Cre:Cre ) was crossed to HIF-1␣-floxed (HIF-1␣ FL/FL ) mice to generate a myeloid-specific HIF-1␣-deficient mouse line. Similarly, KLF2-floxed (KLF2 FL/FL ) mice were crossed with LysM Cre:Cre mice to generate myeloid-specific KLF2-deficient mice. All mouse colonies were maintained in a clean animal facility, and all animal experimentation was approved by the Case Western Reserve University Institutional Animal Care and Use Committee.
Mice (8 -12 weeks old) were injected intraperitoneally with a mixture of lipoteichoic acid (3 mg/kg) and peptidoglycan (1 mg/kg) or saline solution. Mice were monitored for 4 days following this injection and their rectal temperature, blood pressure, and heart rate were recorded. The shock index was calculated using the following formula: Shock index ϭ heart rate/ systolic blood pressure. Survival data were analyzed by the construction of Kaplan-Meier plots and use of the log-rank test.
Real-time Quantitative RT-PCR, Luciferase Reporter Assay, and Western Blot Analysis-Total RNA was isolated from RAW 264.7 cells or mouse primary peritoneal macrophages and neutrophils following indicated treatment using TRIzol reagent (Invitrogen). 1 g of total RNA was reverse-transcribed using M-MuLV reverse transcriptase in the presence of random hexamers and oligo(dT) primers. Real-time quantitative PCR was performed using Universal SYBR Green PCR Master Mix on an Applied Biosystems Step One plus Real-Time PCR system by applying gene-specific primers.
A luciferase reporter plasmid driven by three tandem copies of the HRE sequence were transfected or cotransfected with KLF2 plasmids in RAW264.7 cells using Lipofectamine 2000 reagent (Invitrogen). Twenty-four hours after transfection, these cells were exposed to live (5 m.o.i.) or heat-inactivated bacteria or bacterial endotoxins and luciferase activity was measured with a luciferase reporter assay system (Promega, Madison, WI). Results are presented as relative luciferase activity over the control group.
Primary mouse peritoneal macrophages or RAW264.7 cells were lysed using radioimmunoprecipitation lysis buffer (Sigma) supplemented with a protease inhibitor mixture tablet (Roche) following the indicated treatment. Equal quantities of total protein were separated by SDS-PAGE and detected by the indicated antibody by immunoblotting assay.
Cell Migration and Invasion Assay-Mouse primary peritoneal macrophages from the indicated genotype or RAW264.7 cells infected with Ad-GFP or Ad-KLF2 were induced with teichoic acid or lipoteichoic acid for 4 h. A cell suspension containing 2 ϫ 10 5 cells was added to the upper chamber of the transwell inserts. The lower chambers of the transwell plates were filled with DMEM supplemented with 5% FBS. These transwell chambers were incubated at 37°C for 18 h in a humidified incubator with an atmosphere of 5% CO 2 and 95% air. Following incubation, cells on the upper part of the wells were removed, and the migrated cells in the lower side of the filter were fixed and stained with Giemsa. The migrated cells on the filter were counted under an inverted microscope. To analyze macrophage-invasive properties, cells treated with teichoic acid or lipoteichoic acid were added to the upper wells of the transwell inserts coated with growth factor-reduced Martigel, and was performed as described previously (24).
Antibiotic Protection Assay-Antibiotic protection assays were performed as described previously (16). Briefly, S. aureus bacteria were grown in logarithmic phase in Luria broth medium. These bacterial cultures were centrifuged to sediment actively proliferating bacteria. These bacterial pellets were washed in sterile ice-cold 1ϫ PBS and diluted with DMEM supplemented with 0.1% BSA to the required concentration. These actively proliferating bacteria were added to the top of the monolayer macrophage culture for 2 h. These cells were washed three times with 1ϫ PBS and incubated with antibioticcontaining media for 20 min to eliminate cell surface-associated bacteria. These macrophage cells were cultured for an additional 6 h and lysed with non-ionic detergent-containing buffer. These macrophage cell lysate-containing intracellular proliferating bacteria were serially diluted with sterile 1ϫ PBS and spread on agar plates to enumerate bacterial colony-forming units.
Statistical Analysis-All data are presented as mean Ϯ S.D. unless indicated otherwise. The statistical significance of differences between two groups was analyzed with Student's t test. Values of p Ͻ 0.05 were considered statistically significant.

Gram-positive Bacteria Wall Components Transcriptionally
Induce HIF-1␣ Expression-Previous studies from our group (24) and others (27,28) indicated that Gram-negative bacterial cell wall components such as lipopolysaccharide induce HIF-1␣ expression in myeloid cells. To determine whether a parallel effect occurred with Gram-positive wall components, we examined whether HIF-1␣ transcriptional activity is modulated by exposure of macrophages to Gram-positive bacteria such as S. aureus. We found that transcriptional activity of HIF-1 was significantly elevated in macrophages following exposure to live (5 m.o.i.) or heat-inactivated, Gram-positive bacteria under normoxic condition (Fig. 1A). Indeed, induction of HIF-1 transcriptional activity by Gram-positive bacteria was as strong as Gram-negative bacteria (E. coli). Intriguingly, individual endotoxin components of Gram-positive bacteria (TA and LTA) robustly induced HIF-1-dependent transcriptional activity in the macrophage cell line (Fig. 1B). Taken together, these results indicated that Gram-positive bacteria and endotoxins derived from Gram-positive bacteria strongly activate HIF-1-dependent transcriptional activity in myeloid cells under normoxic condition.
We next sought to determine whether the induction of HIF-1 transcriptional activity by Gram-positive bacteria and their endotoxins led to an equitable increase in HIF-1␣ mRNA and protein expression. Accordingly, wild-type mouse peritoneal macrophages were exposed to live (5 m.o.i.) or heat-inactivated E. coli or S. aureus and expression of HIF-1␣ mRNA and protein were analyzed by quantitative PCR and immunoblot analysis. As shown in Fig. 1C, both live and heat-inactivated S. aureus induced HIF-1␣ mRNA expression. Importantly, inoculation of live or heat-inactivated S. aureus with wild-type macrophages also robustly induced HIF-1␣ at the protein level (Fig. 1E). Taken together, these results indicate that the heatstable component of S. aureus induced HIF-1␣ expression at mRNA and protein levels in macrophages. Therefore, we next evaluated the individual endotoxin component of Gram-positive bacteria such as TA and LTA on expression of HIF-1␣ mRNA and protein in wild-type mouse peritoneal macrophages. As shown in Fig. 1, D and F, compared to control treatment, both TA and LTA induced HIF-1␣ mRNA and protein expression.
Kinetics of Teichoic and Lipoteichoic Acid Induced HIF-1␣ Expression-Our previous experiments indicated that Gram-positive endotoxins induce HIF-1␣ at mRNA and protein levels. Next, we analyzed the kinetics of HIF-1␣ mRNA and protein induction by TA in wild-type mouse peritoneal macrophages. Teichoic acid induced HIF-1␣ mRNA expression in a dose-dependent manner ( Fig. 2A) up to 4 g/ml. Interestingly, a higher dose (8 -10 g/ml) of TA did not alter HIF-1␣ expression at both mRNA and protein levels. Furthermore, the time-dependent HIF-1␣ mRNA and protein expression analysis indicated that TA (4 g/ml) induced HIF-1␣ mRNA and protein expression as early as 6 h, an effect that was sustained for up to 24 h (Fig. 2B). A similar pattern of HIF-1␣ mRNA and protein expression was observed in wild-type mouse peritoneal macrophages in response to LTA (Fig. 2, C and D). Collectively, these results indicate that Gram-positive endotoxins such as TA and LTA induce HIF-1␣ mRNA and protein expression in a dose-and time-dependent manner.
Gram-positive Endotoxins Induced Myeloid Inflammatory Gene Expression in a HIF-1␣-dependent Manner-The studies above identify HIF-1␣ as important in the response to Gram-positive infection. We next sought to examine whether Grampositive bacterial products affected key myeloid cellular processes. Accordingly, mouse peritoneal macrophages derived from control (LysM Cre/Cre ) and HIF-1␣ ⌬/⌬ mice were stimulated with TA or LTA and subjected to a cell migration or inva- with teichoic acid (4 g/ml), lipoteichoic acid (4 g/ml), and LPS (100 ng/ml) for 12 h. HRE-luciferase activity was measured and indicated as relative fold changes over control. C, wild-type mouse primary peritoneal macrophages were exposed with live (5 m.o.i.) or heat-inactivated E. coli and S. aureus for 6 h. D, wild-type mouse peritoneal macrophages were induced with LPS (100 ng/ml), teichoic acid (4 g/ml), and lipoteichoic acid (4 g/ml) for 6 h. Total RNA was isolated, and HIF-1␣ mRNA expression was analyzed by quantitative PCR and normalized to 36B4. E, wild-type mouse primary peritoneal macrophages were incubated with live (5 m.o.i.) heat-inactivated E. coli and S. aureus for 6 h. F, wild-type mouse peritoneal macrophages were induced with LPS (100 ng/ml), teichoic acid (4 g/ml), and lipoteichoic acid (4 g/ml) for 6 h. Cell lysates were analyzed for HIF-1␣ protein expression by immunoblot analysis. MDA-MB-231 cells exposed to 6 h of hypoxia were used as a positive control, and actin was used as a loading control. Combined data of three experiments are shown in each case. Data represent mean Ϯ S.D. *, p Ͻ 0.05 versus the indicated control. JANUARY 6, 2012 • VOLUME 287 • NUMBER 2 JOURNAL OF BIOLOGICAL CHEMISTRY 1451 sion assay (Fig. 3, A and B). Although both agents robustly induced macrophage cell migration and invasion, this effect was significantly attenuated in HIF-1␣-deficient macrophages. To further evaluate the functional consequences of HIF-1␣ activation by Gram-positive bacteria and TA/LTA, we performed an antibiotic protection assay to enumerate intracellular killing of S. aureus by control (LysM Cre/Cre ) macrophages compared with intracellular bacterial killing by those derived from the HIF-1␣-deficient macrophages (Fig. 3C). Our results indicate that deficiency of HIF-1␣ significantly attenuates macrophage intracellular bacterial killing and results in a significantly higher number of colony-forming units in an antibiotic protection assay.

KLF2 Regulates Gram-positive Toxic Shock Syndrome
We next sought to determine whether the altered cellular response to Gram-positive toxins in HIF-1␣-deficient macrophages was due to diminished expression of HIF-1␣ target genes that modulate these cellular processes. Accordingly, control (LysM Cre/Cre ) and HIF-1␣ ⌬/⌬ mice peritoneal macrophages and neutrophils were stimulated with TA or LTA, and expres-sion of HIF-1␣ target genes that are involved in intracellular bacterial killing and cellular motility were analyzed (Fig. 3,  D-H). Both TA and LTA robustly induced iNOS expression in control macrophages, an effect that was significantly attenuated in HIF-1␣-deficient macrophages (Fig. 3D). This reduction in iNOS expression in HIF-1␣-deficient macrophages corresponded with a decrease in macrophage intracellular bacterial killing ability. Further, TA and LTA treatment increased expression of COX-2, MMP-2, IL-1␤, and IL-6 in macrophages and neutrophils derived from control mice; effects that were significantly diminished in HIF-1␣-deficient myeloid cells (Fig. 3, E-H). Taken together, these results indicated that Gram-positive bacterial endotoxins mediate some of their key effects through activation of the HIF-1␣ signaling pathway.
KLF2 Modulates Gram-positive Bacterial Endotoxin-induced HIF-1␣ Expression-Previous work from our group and others has implicated KLF2 as a critical transcriptional regulator of myeloid cell activation and HIF-1␣ expression (23, 24,  A and B, upper panels). HIF-1␣ protein levels were analyzed by immunoblot analysis (lower panels). Wild-type mouse primary peritoneal macrophages were stimulated with 0 -8 g/ml lipoteichoic acid (C) for 0 -24 h (D). HIF-1␣ mRNA expression was analyzed by quantitative PCR and normalized to 36B4 (C and D, upper panels). HIF-1␣ protein levels are analyzed by immunoblot (lower panels). MDA-MB-231 cells exposed to hypoxia were used as a positive control, and actin was used as a loading control. Combined data of three experiments is shown in each case. Data represent mean Ϯ S.D. 29). Therefore, as a first step, we examined whether Grampositive bacteria and bacterial products modulate KLF2 expression in macrophages. Accordingly, wild-type mouse peritoneal macrophages were exposed to heat-inactivated S. aureus, TA, and LTA separately, and expression of KLF2 at the mRNA and protein levels was examined by quantitative PCR and immunoblot analysis respectively (Fig. 4, A and B). Exposure of wildtype macrophages to heat-inactivated S. aureus, TA, or LTA significantly reduced KLF2 expression at both the mRNA and protein levels. Next, we evaluated whether KLF2 regulates the Gram-positive bacterial product-induced HIF-1 transcriptional activity in macrophages (Fig. 4C). Overexpression of KLF2 alone significantly attenuated TA-or LTA-induced HIF-1 transcriptional activity in RAW264.7 cells. Next, RAW264.7 cells were infected with Ad-GFP/Ad-KLF2 or peritoneal macrophages from control (LysM cre:cre ) and KLF2-deficient (KLF2 ⌬/⌬ ) mice were stimulated with TA/LTA. The expression of HIF-1␣ at the mRNA and protein levels was ana-lyzed by quantitative PCR and immunoblot analysis respectively (Fig. 4, D-G). Overexpression of KLF2 significantly reduced TA-or LTA-induced HIF-1␣ mRNA and protein expression in RAW264.7 macrophage cells (Fig. 4, D and F). Conversely, deficiency of KLF2 significantly elevated HIF-1␣ mRNA and protein expression in KLF2 ⌬/⌬ mouse macrophages (Fig. 4, E and G). Collectively, these data suggest that exposure to Gram-positive products confers an anti-parallel effect on KLF2 and HIF-1␣ expression.

KLF2 Regulates Gram-positive Toxic Shock Syndrome
KLF2 Suppresses TA/LTA-induced Macrophage Cell Functions and Gene Expression-Our results suggest that TA or LTA facilitate inflammatory gene expression and cellular functions through a dual mechanism-suppression of KLF2 expression and activation of the HIF-1␣ signaling pathway. This prompted us to examine whether KLF2 modulates TA-or LTA-induced HIF-1␣-dependent inflammatory gene expression and cellular functions in macrophages. Accordingly, RAW264.7 cells overexpressing Ad-KLF2 or peritoneal macro-FIGURE 3. Gram-positive, endotoxin-induced macrophage cell motility, bactericidal function, and inflammatory gene expression are HIF-1␣-dependent. A and B, primary peritoneal macrophages from control (LysM Cre/Cre ) and HIF-1␣ ⌬/⌬ mice were stimulated with 4 g/ml teichoic or lipoteichoic acid and subjected to a migration or invasion assay. The number of cells migrated (A) or invaded (B) in unstimulated wells across the membrane were assigned as 100%, and fold changes over this are indicated. C, primary peritoneal macrophages from control (LysM Cre/Cre ) and HIF-1␣ ⌬/⌬ mice were inoculated with S. aureus, and intracellular bacterial killing was analyzed by antibiotic protection assay. D-H, primary peritoneal macrophages and neutrophils from control (LysM Cre/Cre ) and HIF-1␣ ⌬/⌬ mice were stimulated with 4 g/ml teichoic or lipoteichoic acid for 6 h. Indicated target genes were analyzed by quantitative PCR and normalized to 36B4. Combined data of three experiments are shown in each case. Data represent mean Ϯ S.D. *, p Ͻ 0.05 versus indicated control.
phages from control (LysM cre:cre ) and KLF2-deficient (KLF2 ⌬/⌬ ) mice were stimulated with TA or LTA and subjected to a cell migration assay (Fig. 5, A and B). Overexpression of KLF2 reduced and deficiency of KLF2 enhanced TA-or LTAinduced macrophage cell migration. Analysis of intracellular bacterial-killing ability of control (LysM cre:cre ) and KLF2-deficient (KLF2 ⌬/⌬ ) macrophages indicated that deficiency of KLF2 enhanced macrophage intracellular bacterial killing (Fig. 5C). Further, inflammatory gene expression analysis indicated that deficiency of KLF2 significantly enhanced Gram-positive, bacterial product-induced expression of iNOS in primary macrophages (Fig. 5D). Concordant with this observation, overexpression of KLF2 suppressed and deficiency of KLF2 induced expression of IL-1␤ and COX2 in macrophages exposed to Gram-positive endotoxins (Fig. 5, E and F). Taken together, these results are consistent with increased expression of HIF-1␣ in KLF2-deficient macrophages (Fig. 4, D-G) and decreased cellular motility, intracellular bacterial killing, and inflammatory gene expression observed in HIF-1␣-deficient macrophages (Fig. 3, A-H).
HIF-1␣ and KLF2 Regulate the Sepsis Phenotype Induced by LTA/TA in Vivo-We observed that Gram-positive bacterial endotoxins such as LTA/TA reduced KLF2 expression and induced HIF-1␣ activation and proinflammatory gene expression in myeloid cells. Given that the elaboration of excess proinflammatory factors can contribute to septic physiology, we hypothesized that HIF-1␣ deficiency may reduce and KLF2 deficiency may enhance susceptibility to sepsis. To test this Wild-type mouse peritoneal macrophages were induced with heatinactivated S. aureus, teichoic acid, and lipoteichoic acid for 6 h. A, KLF2 expression was analyzed by quantitative PCR and normalized to 36B4. B, KLF2 protein levels were analyzed by immunoblot analysis. Actin was used as a loading control. C, RAW264.7 cells were cotransfected with a HRE-luciferase reporter plasmid and a pcDNA3 or KLF2 overexpression plasmid. These cells were stimulated with 4 g/ml teichoic acid or lipoteichoic acid for 12 h. HRE-luciferase activity was measured and indicated as relative fold changes over control. D-G, RAW264.7 cells infected with Ad-GFP/Ad-KLF2 (D and F) or primary peritoneal macrophages from control (LysM Cre/Cre ) and KLF2 ⌬/⌬ mice (E and G) were induced with 4 g/ml teichoic acid or lipoteichoic acid for 6 h. HIF-1␣ mRNA and protein levels were analyzed by quantitative PCR and normalized to 36B4. HIF-1␣ protein expression was analyzed by immunoblot analysis. Combined data of three experiments are shown in each case. Data represent mean Ϯ S.D. *, p Ͻ 0.05 versus indicated control. hypothesis, we examined whether myeloid KLF2 or HIF-1␣ deficiency modulates LTA-induced sepsis syndrome in vivo. Control (LysM Cre/Cre ), KLF2 ⌬/⌬ , and HIF-1␣ ⌬/⌬ mice were intraperitoneally injected with a combination of LTA and PGN. This challenge in KLF2 ⌬/⌬ mice induced 100% mortality by 72 h, whereas control (LysM Cre/Cre ) mice experienced only 50% mortality. By contrast, myeloid deficiency of HIF-1␣ was protective against Gram-positive toxin-induced mortality (Fig.  6A). In addition, although myeloid-specific KLF2-deficient mice exhibited all the cardinal features of endotoxic shock, including hypothermia, hypotension, and elevated shock index (Fig. 6B-D), the myeloid-specific HIF-1␣-deficient mice were protected from these effects. These in vivo observations are consistent with our in vitro studies that the Gram-positive bacterial endotoxins mediate their proinflammatory effects through modulating KLF2 and HIF-1␣ expression in myeloid cells.

DISCUSSION
The central findings of this study are that 1) Gram-positive bacterial endotoxins transcriptionally induce HIF-1␣ expression; 2) teichoic and lipoteichoic acid induced HIF-1␣ expression in a dose-and time-dependent manner; 3) Gram-positive endotoxins induced macrophage cell motility, bacterial killing, and inflammatory gene expression in an HIF-1␣-dependent manner; 4) KLF2 modulates Gram-positive bacterial endotoxin-induced HIF-1␣ expression; 5) KLF2 suppresses TA/LTAinduced macrophage cellular functions and gene expression; and 6) HIF-1␣ and KLF2 regulate the sepsis phenotype induced by LTA/TA in vivo. Collectively, these observations identify HIF-1␣ as central regulator of Gram-positive bacterial endotoxin-mediated sepsis and that KLF2 modulates this pathophysiological response by regulating expression of HIF-1␣ in myeloid cells.
The myeloid cell response to infection is a central component of host defense. These cells invade infected tissues to contain and combat the invading organism. If containment fails and bacteria/bacterial products leach out into the circulation, a vicious self-perpetuating cycle of inflammation can lead to septic shock (30). Our study provides insights regarding the molecular basis for these events in the context of Gram-positive infection. We note that tissue foci of infection are characterized by very low levels of oxygen and glucose in combination with high concentrations of lactate and free oxygen radicals. Therefore, the innate immune cells that respond to infection must adapt to these adverse condition quickly to combat the invading organism (31). Our studies suggest that Gram-positive bacteria and endotoxins induce HIF-1␣ mRNA, protein expression, and transcriptional activity. This increase in HIF-1 signaling is beneficial to myeloid cell adaptation to the hypoxic condition by increasing the rate of glycolysis and offering protection from apoptosis (32). In addition, increased HIF-1 also enhances the production of antimicrobial agents such as iNOS from myeloid cells that may contribute to antimicrobial properties of myeloid cells. The importance of HIF-1 in this response to Gram-positive infection is strongly supported by the observation that myeloid-specific deficiency of HIF-1 abrogates many of the cardinal cellular responses to infection including migration, inva- A and B, RAW264.7 cells infected with Ad-GFP/Ad-KLF2 or primary peritoneal macrophages derived from control (LysM Cre/Cre ) and KLF2 ⌬/⌬ mice were stimulated with 4 g/ml teichoic or lipoteichoic acid and subjected to a migration assay. The number of cells migrated in unstimulated wells across the membrane was assigned as 100%, and fold changes over this are indicated. C, primary peritoneal macrophages from control (LysM Cre/Cre ) and KLF2 ⌬/⌬ mice were incubated with S. aureus, and the intracellular bacterial-killing ability of these macrophages was analyzed by antibiotic protection assay. D-F, RAW264.7 cells infected with Ad-GFP/Ad-KLF2 or primary peritoneal macrophages from control (LysM Cre/Cre ) and KLF2 ⌬/⌬ mice were stimulated with 4 g/ml teichoic or lipoteichoic acid for 8 h. Indicated target gene expression was analyzed by quantitative PCR and normalized to 36B4. Combined data of three experiments are shown in each case. Data represent mean Ϯ S.D. *, p Ͻ 0.05 versus the indicated control. JANUARY 6, 2012 • VOLUME 287 • NUMBER 2 sion, bacterial killing, and cytokine production. However, this response is truly a double-edged sword because if this induction of HIF-1 is not carefully controlled, excessive myeloid cell activation can ensue and lead to a systemic cytokine storm and endotoxic shock. Indeed, previous studies indicate that Grampositive bacterial endotoxins can induce shock and multiple organ failure (10,11). Consistent with this observation, we also observed dramatic mortality in response to LTA. Importantly, this effect on mortality was strongly attenuated in the absence of HIF-1. These findings are reminiscent of the protective effect of HIF-1 deficiency observed by Peyssoneux et al. (18) in the context of Gram-negative sepsis.

KLF2 Regulates Gram-positive Toxic Shock Syndrome
Given the importance of carefully titrating HIF-1 activation, stringent regulatory mechanisms must clearly exist. In this regard, the identification of KLF2 is particularly noteworthy. Previous studies from our group were first to implicate KLF2 as an inhibitor of myeloid proinflammatory activation (23,24). More recently, we showed that myeloid-specific deletion of KLF2 led to spontaneous low-level myeloid activation even under basal conditions (24). Following stimulation with agents such as LPS, KLF2-deficient macrophages exhibited hyperinduction of HIF-1 and numerous cytokines/chemokines. In this work, we found that the same stimuli (LTA/TA) that induced HIF-1 also reduced KLF2 expression. This reduction of KLF2 is likely important, as it allows for HIF-1 induction and bactericidal activity. Consistent with this idea, overexpression of KLF2 inhibited and deficiency led to an exaggerated induction of HIF-1. Our in vivo studies also provide support for the importance of this KLF2-HIF-1 regulatory axis in vivo. Given that unbridled HIF-1 activity can be deleterious to the host, we observed that myeloid-KLF2-null mice succumbed more readily to LTA-induced sepsis.
We note that a previous study from Schwander et al. (14) indicated that lipoteichoic acid mediate their cellular effect through binding Toll-like receptor 2 (TLR2) and induction of NF-B transcriptional activity. A recent study from our group and others indicated that Gram-negative endotoxins (e.g. LPS) bind TLR4 and induce HIF-1␣ expression through transcriptional activation of NFB (24,28). Given these observations, it is likely that a parallel mechanism accounts for the TA/LTAmediated induction of HIF-1. Intriguingly, NFB activation has also been shown to inhibit KLF2 expression. Thus, the initial activation of NFB by Gram-positive and Gram-negative endotoxins may account for the parallel induction of HIF-1 and reduction of KLF2. These observations also provide insights as to how KLF2 may prevent the transcriptional induction of HIF-1. Initial studies from our group showed that KLF2 can inhibit NFB transcriptional activity in numerous cell types, including myeloid cells (23). More recently, we determined that this inhibition occurs through sequestration of key coactivators (e.g. p300/pCAF) required for optimal NF-B transcriptional activity (24). Thus, a mutually antagonistic relationship exists between KLF2 and NFB, and the balance of these two factors is ultimately important in determining HIF-1 levels and attendant downstream events.
In sum, the in vitro and in vivo observations presented here underscore the importance of a KLF2-HIF-1 pathway in regulating myeloid cell function in the context of Gram-positive FIGURE 6. HIF-1␣ and KLF2 regulate the Gram-positive, endotoxin-mediated sepsis phenotype in vivo. A, age-and sex-matched control (LysM Cre/Cre ), HIF-1␣ ⌬/⌬ , and KLF2 ⌬/⌬ mice were challenged with lipoteichoic acid (3 mg/kg) supplemented with peptidoglycan by intraperitoneal injection. These mice were observed for 96 h for survival. B-D, age-and sex-matched control (LysM Cre/Cre ), HIF-1␣ ⌬/⌬ , and KLF2 ⌬/⌬ mice were challenged with lipoteichoic acid (3 mg/kg) containing peptidoglycan by intraperitoneal injection and were monitored for changes in core body temperature (rectal probe), systolic blood pressure (tail cuff blood pressure monitor), and shock index (shock index ϭ heart rate/systolic blood pressure). n ϭ 10 mice per group in each experiment (A-D). Data represent mean Ϯ S.D. *, p Ͻ 0.05 versus the indicated control.
sepsis. These results, coupled with our recent findings in Gramnegative infections, identify this pathway as a major regulator of myeloid function in host defense. Consequently, agents targeting this pathway may allow for therapeutic gain in the treatment of bacterial infections and sepsis.