Hypoxia-Inducible Factor (HIF) 1α Accumulation and HIF Target Gene Expression Are Impaired after Induction of Endotoxin Tolerance¹

Bacterial LPS, also known as endotoxins, are the most important cell wall components of Gram-negative bacteria and elicit a systemic inflammatory response (1). LPS activate leukocytes to produce inflammatory cytokines including TNF-α and IL-1β. LPS-induced proinflammatory molecules are indispensable for counteracting the growth and dissemination of Gram-negative bacteria, but overproduction can cause a lethal endotoxin shock. LPS bind to a receptor complex involving membrane-bound CD14 and TLR4, a distinct member of the TLR family with ligand specificity for LPS (1). TLRs are evolutionary conserved proteins that recognize pathogen-associated molecular patterns produced by bacteria or other pathogens and mammalian TLRs are homologous to members of the IL-1R family (2, 3).

An initial exposure to a low dose of LPS induces a transient state of hyporesponsiveness to a subsequent challenge with LPS (4). The detailed mechanisms underlying this phenomenon known as endotoxin tolerance remain to be resolved and the contribution of several LPS signaling components including the NF-κB transcription factor family has been discussed controversially. However, it is commonly accepted that endotoxin tolerance leads to an impaired release of proinflammatory cytokines like IL-1β and TNF-α from monocytes and macrophages after LPS stimulation (1). These cytokines have been shown to induce the accumulation and activation of the transcription factor complex hypoxia-inducible factor 1 (HIF-1)³ independently of hypoxia (5, 6, 7). HIF-1 belongs to a family of heterodimeric transcription factors that regulates the expression of genes involved in angiogenesis, oxygen transport, glucose metabolism, and vascular tone (8). HIF-1 is formed by two subunits: HIF-1β, which is constitutively expressed, and HIF-1α, which is regulated by cellular oxygen tension. Under normoxic conditions, the HIF-1α protein becomes hydroxylated on specific proline residues and binds to the von Hippel-Lindau protein, which targets HIF-1α for ubiquitinylation and proteasomal degradation. Under hypoxic conditions, prolyl hydroxylation is reduced, thus finally leading to an increase in HIF-1α protein and expression of HIF-1 target genes (9). In addition to hypoxia, which is also a typical feature of the tissue microenvironment during bacterial infection, multiple inflammatory stimuli lead to HIF-1α accumulation and activation (10, 11). Along with this, an important role of HIF-1 for the host immune response during bacterial infection has recently been shown (12, 13). We have previously identified LPS itself as potent inducer of HIF-1α accumulation, primarily by a NF-κB-dependent activation of HIF-1α gene expression (7). All of these facts indicate an integrative role for HIF-1 in conditions of hypoxia and inflammation (14).

With respect to systemic inflammation and septic shock, experimental evidence suggests that endotoxin tolerance serves to protect the organism from a second potentially lethal endotoxin dose during inflammation and sepsis (15). In contrast, survivors of septic shocks have an increased incidence of subsequent bacterial infections due to a suppressed monocyte/macrophage response to LPS (16).

Because HIF-1 and HIF-1 target genes play a pivotal role during hypoxia and inflammation, we raised the question whether induction of endotoxin tolerance influences the ability of monocytic cells to cope with hypoxia. We investigated HIF-1 accumulation, expression of HIF-1 target genes, and cellular characteristics under hypoxic conditions after induction of endotoxin tolerance in the human monocytic cell line THP-1 and in primary mouse peritoneal macrophages. We further analyzed the molecular mechanisms underlying endotoxin tolerance with respect to hypoxic as well as LPS-induced HIF-1 activation and defined a distinct role for the RelB/p100/p52 members of the NF-κB family.

Bacterial LPS (LPS serotype 0111:B4) was obtained from Sigma-Aldrich. PMA was obtained from Calbiochem. The concentrations of all reagents used in the experiments were not toxic for the cells as judged from a MTT assay (17). The protease inhibitor mixture was from Roche Diagnostics.

RPMI 1640 cell culture medium was from BioWhittaker or PAA and FBS from Biochrom. All other cell culture supplements were obtained from Invitrogen. Small interfering RNA (siRNA) against RelB/p100 and nontarget siRNA were obtained from Dharmacon (order no. 3918-05). Matrigel was obtained from BD Biosciences.

Ten- to 12-wk-old C57BL/6 mice were pretreated with 240 μg of LPS dissolved in PBS by i.p. injection (18). Control mice only received PBS. Four days later, mice were sacrificed and peritoneal macrophages were isolated by adding 5 ml of RPMI 1640 medium containing Liquemin (5 U/ml) to the peritoneal cavity. The medium with macrophages was aspired, cells were counted, resuspended in fresh medium (RPMI 1640, very low endotoxin (PAA) supplemented with 2% FBS, 2 mM l-glutamine, 1 mM sodium pyruvate, 100 U/ml penicillin, and 100 μg/ml streptomycin), and seeded in 6-well culture plates (2 × 10⁶ cells/well). Cells were allowed to adhere to the culture plate before starting the experiments.

The human monocytic cell line THP-1 was obtained from the German Collection for Microorganisms and Cells. THP-1 cells were cultured in RPMI 1640 medium, supplemented with 10% FBS, 1.5 mM l-glutamine, 1 mM sodium pyruvate, 100 U/ml penicillin, and 100 μg/ml streptomycin in a humidified atmosphere of 5% CO₂ in air. Cells were split once per week to maintain a cell density of 10⁶cells/ml. For all experiments, 2-ml cell suspensions (10⁶cells/ml) were used in 35-mm petri dishes. Cells were treated with 10 nM PMA for 4 days, resulting in adherence and differentiation into a macrophage-like phenotype. Differentiation was confirmed as described previously (7). Hypoxic conditions were achieved by placing the culture dishes in a Heraeus incubator at different oxygen concentrations (5, 3, and 1% O₂) with 5% CO₂ and N₂ as balance. LPS was added to the cells as indicated for each experiment. At the end of the experiments, cells were lysed in 4M GTC followed by total RNA extraction for the determination of specific mRNAs by RT-PCR. Additionally, total cell lysates and nuclear extracts were prepared and submitted to immunoblotting and measurement of DNA-binding activity. Low passage number cells were used for all experiments.

Differentiated THP-1 cells were kept in serum-free RPMI 1640 medium only (naive) or made tolerant to endotoxins by pretreating with 0.05 ng/ml LPS for 48 h. After renewal of cell culture medium, naive and tolerant cells were further incubated with or without 1 μg/ml LPS for the indicated time periods and oxygen concentrations.

For siRNA transfection experiments, THP-1 cells were seeded on petri dishes with 10-cm diameter and differentiated as described above. Differentiated THP-1 cells were transfected with 20 nM RelB/p100 siRNA and a nontarget siRNA (luciferase) using Oligofectamine (Invitrogen) before starting the tolerance induction protocol.

Total RNA was isolated by the acidic guanidinium thiocyanate-phenol-chloroform extraction method (19). Total RNA was reverse transcribed into cDNA and quantification of HIF target genes was performed by real-time PCR as previously described (7). The primer sequences used for qualitative and quantitative PCR are listed in Table I. Specificity of primers was checked by nucleotide BLAST and confirmed by sequencing of the PCR amplicons. Amounts of specific cDNA were normalized to β-actin and expression was calculated as relative induction to the respective controls.

For whole cell lysate preparation, cells were lysed with 65 μl of lysis buffer (150 mM NaCl, 10 mM Tris (pH 7.9), 1 mM EDTA, 0.1% Igepal, 1× protease inhibitor mixture; Roche Diagnostics) for 20 min on ice. The lysates were centrifuged at 3600 rpm at 4°C for 5 min in a microcentrifuge; supernatants containing cellular proteins were collected and stored at −80°C. Protein concentrations of the supernatants were quantitated using the Bio-Rad protein assay reagent.

Nuclear proteins were prepared using a NE-PER nuclear extraction kit (Pierce) according to the manufacturer’s instructions. Nuclear proteins were stored at −80°C until use. Protein concentrations were determined using a Bio-Rad kit.

Fifty micrograms of total cell lysate per lane was subjected to 7.5% SDS-PAGE and transferred onto a nitrocellulose membrane (0.2-μm pore size; Schleicher & Schuell Microscience). Human HIF-1α was detected using a mouse mAb against HIF-1α (Transduction Laboratories) and mouse HIF-1α was detected with a rabbit polyclonal anti HIF-1α Ab (Novus Biologicals). α-Tubulin (Santa Cruz Biotechnology) served as a loading control.

Before the addition of whole cell lysate buffer or GTC, respectively, cell-free supernatants were collected by centrifugation, transferred to new tubes, and kept at −80°C until analysis. Il-1β was measured by the Quantikine ELISA system according to the manufacturer’s instructions (R&D Systems).

Five micrograms of nuclear proteins was used to detect the DNA-binding activity of the p65, p50, and p52 members of the NF-κB transcription factor family by an ELISA-based assay according to the manufacturer’s instructions (TransAM NF-κB Family Transcription Factor ELISA; Active Motif).

Viability of naive and tolerant THP-1 was judged from the MTT assay (17).

Liquid Matrigel was 1/3 diluted with serum-free medium and allowed to polymerize overnight on polysterol dishes (250 μl/cm²). One × 10⁶ THP-1 cells, naive or LPS tolerant, were added, stimulated with LPS (1 μg/ml), and incubated for 90 min. After removal of the supernatant containing the noninvaded cells, cells were fixed with 4% paraformaldehyde solution, stained with H33342 (Hoechst), and subjected to microscopic analysis. For each condition, invaded cells were counted in four equal and representative areas (regions of interest).

All densitometric analyses were performed using the ImageJ program (National Center for Biotechnology Information).

All experiments were performed in triplicates. Values of mRNA quantification are means ± SD. Values of cell counts in the invasion assay are given as means ± SD. The Dunnett test and Tukey-Kramer test were used to calculate whether the differences of means between treated and control groups were significant.

THP-1 cells were differentiated with 10 nM PMA for 4 days. Differentiation was confirmed by morphological changes, adherence, and the increased expression of the macrophage differentiation marker CD68 (data not shown). First, inducibility of endotoxin tolerance was tested by pretreatment of differentiated THP-1 cells with 0.05 ng/ml LPS for 48 h. The second stimulation with LPS was done under normoxic (20.9% O₂) and hypoxic (3% O₂) conditions for 6 h. Treatment with LPS lead to an increase in IL-1β mRNA expression and IL-1β protein release. This induction seen in naive THP-1 cells was significantly reduced in endotoxin tolerant cells (Table II). To exclude a direct effect of endotoxin tolerance on the abundance of surface molecules necessary for LPS binding, we screened the expression of TLR4 and CD14 in tolerant THP-1. Both surface molecules were markedly up-regulated in differentiated THP-1, but no significant difference between naive and tolerant cells was found (Fig. 1).

Differentiated THP-1 cells were pretreated with LPS (0.05 ng/ml) for 48 h followed by an acute stimulation with 1 μg/ml LPS for 6 h. Analysis of HIF-1α protein content revealed a significant reduction of LPS-induced HIF-1α accumulation in LPS-tolerant THP-1 cells under normoxic conditions. Next, we raised the question whether also hypoxia-induced HIF-1α accumulation is altered in endotoxin-tolerant cells. THP-1 cells were made tolerant to LPS before exposure to different degrees of hypoxia (5, 3, and 1% oxygen) in the presence or absence of LPS. In tolerant cells, hypoxia-induced HIF-1α protein accumulation was significantly lower. In addition, the LPS induced HIF-1α accumulation was reduced (Fig. 2,A). Since LPS is believed to increase HIF-1α predominantly at the transcriptional level, we analyzed the expression of HIF-1α mRNA in tolerant cells. Constitutive HIF-1α mRNA levels were significantly reduced upon induction of endotoxin tolerance in normoxia. After 6 h under hypoxic conditions, the constitutive HIF-1α mRNA expression in naive cells was diminished independently of the degree of hypoxia. In tolerant cells, the expression level of HIF-1α was further diminished ∼50% when compared with naive hypoxic cells. As expected, the LPS-induced HIF-1α expression was significantly reduced in tolerant THP-1 cells independently from oxygen tension (Fig. 2 B).

Reduced accumulation of HIF-1α protein could be due to diminished transcription and translation as well as increased degradation. We and others (7, 20) have previously shown that LPS up-regulate the transcription of the HIF-1α mRNA in a NF-κB-dependent manner. Therefore, we analyzed the DNA-binding capacity of three members of the NF-κB transcription factor family in naive and tolerant cells. LPS-induced a dramatic, ∼20-fold increase in p65 DNA binding in naive THP-1 cells, but only a moderate increase in p50 DNA binding. Induction of endotoxin tolerance alone evoked a weak stimulation of p65 and p50 DNA binding. Whereas the LPS-induced p65 DNA binding in tolerant cells was diminished, ∼50% when compared with control, the induction of endotoxin tolerance had no influence on the LPS-induced DNA binding of the p50 subunit (Fig. 3,A). DNA binding of p52 was not influenced by acute LPS exposure but was significantly up-regulated in tolerant THP-1 (Fig. 3 B).

The p52 member of the NF-κB transcription factor family has already been described to be involved in the induction of endotoxin tolerance. To elucidate the role of activation of RelB/p100/p52 for the induction of tolerance, we knocked down the p52 precursor RelB/p100 by siRNA. In nontarget siRNA-treated cells, endotoxin tolerance was inducible and HIF-1α mRNA expression was decreased. After treatment with RelB/p100 siRNA, the tolerance-induced decrease of HIF-1α expression was abolished and, amazingly, in naive cells a significant increase in HIF-1α expression was observed (Fig. 3,C). The accumulation of HIF-1α protein reflected the changes seen on the HIF-1α mRNA expression level (Fig. 3 D).

Finally, the consequences of endotoxin tolerance for the survival and function of mononuclear cells under hypoxic conditions were investigated. Since macrophages often have to survive and function in a microenvironment with reduced oxygen supply, we first analyzed whether induction of endotoxin tolerance exerts effects on the viability of THP-1 under hypoxic conditions. Tolerant as well as naive cells were cultured for 24 h under normoxic and hypoxic conditions. Viability of THP-1 cells was judged from the MTT assay. Induction of endotoxin tolerance had no effect on cells under normoxic conditions, likewise incubation of naive cells with LPS did not change viability (data not shown). Under hypoxic conditions (3% O₂), the viability of endotoxin-tolerant THP-1 was significantly reduced when compared with the naive control cultures (Fig. 4,A). Survival in a hypoxic microenvironment implicates the generation of energy by glycolysis. Therefore, we quantitated the expression of phosphoglycerate kinase 1 (PGK1), a classical hypoxia-inducible HIF-1 target gene. Naive and tolerant THP-1 cells were exposed to hypoxia (3% O₂) in the presence or absence of LPS for 6 h. LPS treatment in naive cells resulted in a 3-fold induction of PGK1 expression. In tolerant cells, a significant reduction in basal PGK1 expression was observed when compared with the naive controls and the LPS-induced expression of PGK1 was abolished (Fig. 4,B). Adrenomedullin (ADM) is a HIF-1 target gene highly expressed under conditions of hypoxia and sepsis. Constitutive as well as LPS-induced expression of ADM was significantly reduced in endotoxin-tolerant cells (Fig. 4 C).

Migration of mononuclear cells toward inflamed tissue areas is another important step in the host immune defense. To establish a model for invasion and migration of mononuclear cells, naive and tolerant THP-1 cells were cultured on Matrigel with or without further LPS stimulation. After 2 h, LPS induced a 40% increase in invasiveness of naive THP-1 cells. Induction of endotoxin tolerance significantly reduced the constitutive invasion of cells under control conditions. In addition, the increase in invasiveness after LPS stimulation was abolished in tolerant cells (Fig. 4 D).

Primary mouse peritoneal macrophages were used to determine the importance of the above-described findings for an in vivo situation. Peritoneal macrophages from LPS-pretreated mice showed a significantly reduced accumulation of HIF-1α protein (Fig. 5,A) and expression of HIF-1α mRNA when compared with macrophages from control mice (Fig. 5,B). The HIF target genes ADM and GLUT1 were expressed on a significantly lower level (Fig. 5, C and D).

The human monocytic cell line THP-1 is a well-established model for studying the induction of endotoxin tolerance in monocytic cells (21). Endotoxin tolerance has been described in bacterial LPS-pretreated macrophages as the impaired release of proinflammatory cytokines on a subsequent challenge with LPS (22). In accordance with previous findings, we demonstrated that induction of endotoxin tolerance reduced the LPS-induced expression and release of the proinflammatory cytokine IL-1β. Confirming our earlier work, LPS was able to induce the accumulation and activity of the transcription factor complex HIF-1 in monocytic cells (7). The activity of HIF-1 is determined by the abundance of the oxygen-regulated subunit HIF-1α (9). In the presence of molecular oxygen (normoxia), HIF-1α becomes hydroxylated at distinct proline residues, resulting in the recruitment of the E3-ligase von Hippel-Lindau protein to the oxygen-dependent degradation domain. Consecutively, HIF-1α is targeted for proteasomal degradation by ubiquitination (23). In the absence of oxygen, hydroxylation of HIF-1α is inhibited and HIF-1α is no longer degraded and accumulates in the nucleus to regulate the cellular response to hypoxia. Apart from this hypoxic HIF-1 activation, several publications during the last years reported that HIF-1α is regulated independently of hypoxia (6, 11, 24). Especially under conditions of inflammation, increased HIF-1α accumulation followed by activation of HIF-1 target genes under nonhypoxic conditions have been described (25, 26). Depending on the cell type and the stimulus, the increase in HIF-1α seemed to occur on transcriptional, translational, or posttranslational cellular levels (27). HIF-1α accumulation after TNF-α and IL-1β stimulation was found to result from increased translation of the HIF-1α mRNA starting at an internal ribosomal entry site located within the nontranslated region of the first exon of the HIF-1α gene (28, 29). In addition, LPS treatment of monocytic cells increased the expression of the HIF-1α mRNA in a NF-κB-dependent manner (7). By increasing transcription and translation, cells overcome the normoxic degradation of the HIF-1α subunit, which in consequence results in the expression of HIF target genes under normoxic conditions.

Since LPS were shown to increase HIF-1α in monocytic cells, we raised the question whether induction of endotoxin tolerance exerts effects on the LPS induced and the constitutive HIF-1α expression in mononuclear cells. And in fact, in tolerant cells, both the constitutive and the LPS-induced HIF-1α mRNA expression and protein accumulation were reduced. The molecular mechanisms underlying the induction of endotoxin tolerance are controversially discussed. Changes in the expression of LPS receptors could be a likely mechanism to reduce LPS responsiveness. But Calvano et al. (30) demonstrated that the expression of components of the LPS receptors complex consisting of TLR4, CD14, and MD2 in LPS-tolerant monocytes was not significantly different from naive controls (30). In our model, we confirmed the induction of endotoxin tolerance by the diminished expression and release of IL-1β after a second LPS stimulation. Like Calvano et al. (30), we observed no changes in the expression of TLR4 and CD14 in tolerant cells. Therefore, it is more likely that changes in intracellular signal transduction are involved in the induction of endotoxin tolerance.

The activation and dominant function of the NF-κB transcription factor complex for LPS-induced gene expression has been shown by several groups (31, 32). We have previously identified NF-κB as the key transcription factor for the LPS-induced HIF-1α mRNA expression (8). Recently, it was confirmed that NF-κB regulates HIF-1α at the transcriptional level (14). Wedel et al. (33) analyzed the role of different members of the transcription factor NF-κB superfamily for the induction of endotoxin tolerance in more detail. They found an up-regulation of the p52 protein and its precursor p100/RelB after induction of LPS tolerance in a human B cell line. In this model, levels of p65 and p50 were not changed by endotoxin tolerance. Overexpression of RelB in THP-1 cells mimics the phenotype of endotoxin tolerance and vice versa, inhibition of RelB abolished the induction of tolerance (21). We observed a moderate ∼2-fold increase in general DNA binding of p65 and p50 under conditions of endotoxin tolerance and, in agreement with the findings of Yoza et al. (21), a highly significant 4-fold induction of p52 DNA binding. In endotoxin-tolerant THP-1 cells, RelB was found to associate with p65, resulting in a repression of the IL-1β promoter activity (21, 34). To test whether p52 is also involved in the reduced expression of HIF-1α seen in tolerant cells, we knocked down p52 by a siRNA approach targeting the p52 precursor RelB/p100. Knocking down p52 not only restored the LPS-induced but also the constitutive expression of the HIF-1α gene and the accumulation of HIF-1α protein in tolerant cells. Moreover, constitutive HIF-1α mRNA levels were significantly increased after RelB/p100 siRNA treatment. It is therefore presumable that p52 is involved as a modulator in the constitutive regulation of HIF-1α gene expression. The most important finding in this work was the down-regulation of basal HIF-1α expression after induction of endotoxin tolerance. Therefore, it was critical to test whether pretreatment with LPS resulting in endotoxin tolerance changes the ability of monocytic cells to cope with hypoxia. As recently demonstrated in an elegant study, the loss of HIF-1α abolished the cardioprotective effects of ischemia-reperfusion, thus identifying HIF-1 as a central component of the cardioprotection by ischemic preconditioning (35). Macrophages from knockout mice lacking HIF-1α in their myeloid cells express significantly reduced levels of the proinflammatory cytokines TNF-α, IL-1α, IL-1β, and IL-12 and are unable to kill bacteria due to reduced expression of the inducible NO synthase (36, 37). These data illustrate that HIF-1 plays a crucial role for the coordination of the innate immune system during bacterial infection (25). We demonstrated in this study that macrophages from mice pretreated with low doses of LPS displayed a phenotype with reduced HIF-1α gene expression, protein accumulation, and HIF target gene expression. Therefore, in vivo induction of endotoxin tolerance mimics the myeloid HIF-1α knockout. Whether the reduced release of inflammatory cytokines in tolerant cells contributes to the reduction of HIF-1 activity remains unclear. However, since the constitutive HIF-1α gene expression is significantly lowered in tolerant cells, it is likely that the additional loss of cytokine production further aggravates the malfunction of the HIF system.

In cases of inflammation or infection, monocytes extravasate from the blood and migrate into the tissue where they differentiate to macrophages. Especially macrophages have to survive and function in hypoxic areas (38, 39). Therefore, the expression of genes enabling the cells to extravasate from the vasculature and generate energy from glycolysis is indispensable for them (40). A recent study further outlined the differential regulation of HIF-1α in monocytes and macrophages (41). We have demonstrated that induction of endotoxin tolerance reduced the hypoxic expression of the HIF-1 target genes ADM and PGK1 in differentiated macrophage-like THP-1 cells. ADM was shown to be critically involved in the development of vascular leakage during sepsis while PGK1 is necessary for the anaerobic metabolism. Probably due to the reduced ability of anaerobic energy generation, the viability of endotoxin-tolerant THP-1 cells under hypoxia was significantly diminished. Moreover, migration into the extracellular matrix was significantly lower in tolerant than in naive cells. Taken together, our results provide evidence that induction of endotoxin tolerance reduces the ability of mononuclear cells to invade, survive, and function in a hypoxic environment. Our findings may further explain why bacterial infections are increased and persist in patients suffering from sepsis.

We thank Daniela Plitzko (Experimentelle Chirurgie, Universitaetsklinikum Essen, Essen, Germany) for assistance with the murine endotoxin tolerance model.

The authors have no financial conflict of interest.