Febrile-range temperature modifies cytokine gene expression in LPS-stimulated macrophages by differentially modifying NF-κB recruitment to cytokine gene promoters

Abstract

We previously showed that exposure to febrile-range temperatures (FRT, 39.5–40°C) reduces LPS-induced TNF-α expression, in part through the direct interaction of heat shock factor-1 (HSF1) with the TNF-α gene promoter. However, it is not known whether exposure to FRT also modifies more proximal LPS-induced signaling events. Using HSF1-null mice, we confirmed that HSF1 is required for FRT-induced repression of TNF-α in vitro by LPS-stimulated bone marrow-derived macrophages and in vivo in mice challenged intratracheally with LPS. Exposing LPS-stimulated RAW 264.7 mouse macrophages to FRT reduced TNF-α expression while increasing IL-1β expression despite the two genes sharing a common myeloid differentiation protein-88 (MyD88)-dependent pathway. Global activation of the three LPS-induced signaling intermediates that lead to cytokine gene expression, ERK and p38 MAPKs and NF-κB, was not affected by exposing RAW 264.7 cells to FRT as assessed by ERK and p38 phosphorylation and NF-κB in vitro DNA-binding activity and activation of a NF-κB-dependent synthetic promoter. However, chromatin immunoprecipitation (ChIP) analysis demonstrated that exposure to FRT reduced LPS-induced recruitment of NF-κB p65 to the TNF-α promoter while simultaneously increasing its recruitment to the IL-1β promoter. These data suggest that FRT exerts its effects on cytokine gene expression in a gene-specific manner through distal effects on promoter activation rather than proximal receptor activation and signal transduction.

the immunomodulatory effects of fever and febrile-range temperatures (FRT; 39.5°C) have been well categorized and include effects on cytokine expression, antigen presentation, and lymphocyte proliferation (16). We previously demonstrated (8, 9, 18, 19, 35–37) that exposure to FRT reduced LPS-stimulated TNF-α expression in the RAW 264.7 macrophage cell line, murine peritoneal macrophages, liver slices, Kupffer cells, human monocytes and monocyte-derived macrophages, and the human THP-1 promonocytic leukemia cell line. In subsequent studies (35–37), we showed that exposure of RAW 264.7 cells to FRT stimulates activation of the stress-activated transcription factor heat shock factor-1 (HSF1), which binds to heat shock response element (HSE)-like sequences present in the TNF-α promoter and represses TNF-α transcription. The role of HSF1 in TNF-α expression was further substantiated by Xiao et al. (44), who showed enhanced LPS-induced TNF-α secretion in HSF1-knockout mice.

Like TNF-α, IL-1β expression is also modified by exposure to hyperthermia. Our group (9) showed that exposing human monocytes and monocyte-derived macrophages to FRT repressed LPS-induced IL-1β expression. In contrast, intratracheal LPS-challenged mice exposed to hyperthermia showed elevated levels of IL-1β in their bronchalveolar lavage fluids (29). Other groups showed that exposure to heat shock (HS) attenuated LPS-induced IL-1β expression in human macrophages (1) and attributed this effect to the physical interaction between HSF1 and C/EBPβ, which each bind to adjacent elements in the human IL-1β promoter (45). In addition, Goldring et al. (13) reported that inducible nitric oxide synthase (iNOS) gene activation was enhanced by coexposure to LPS and HS and that the effect was mediated through HSF1 interactions with the iNOS gene promoter. We reported (34) that coexposure to HS enhanced TNF-α- and IL-1β-induced IL-8 expression in human lung epithelium-like A549 cells through a mechanism that required HSF1 and its interaction with HS response elements present in the IL-8 gene promoter. Moreover, Inouye et al. (17) reported that HSF1, in unstressed cells, regulated LPS-induced IL-6 gene expression by modifying the chromatin accessibility of other transcription factors. These studies clearly demonstrate that hyperthermia in the febrile to HS range can profoundly alter the expression of various LPS-induced genes and that the effect is mediated, at least in part, through HSF1. However, it is not clear how hyperthermia selectively modifies the expression of these LPS-induced genes and whether exposure to hyperthermia also modifies the LPS-induced upstream signaling events leading to gene activation.

LPS stimulates proinflammatory gene expression by engaging the Toll-like receptor (TLR)4 complex that activates bifurcating signaling pathways leading to activation of NF-κB and the MAPK cascades (5, 21, 43). One signaling pathway utilizes the TLR4 adapter molecules myeloid differentiation protein-88 (MyD88) and TIRAP and induces rapid TNF-α and IL-1β gene activation, whereas the second pathway utilizes the adapter molecules TRIF and TRAM to activate IFN-β and regulated upon activation, normal T cell expressed and secreted (RANTES) gene expression (11). The objectives of this study were to determine how exposure to FRT selectively modifies LPS-induced gene expression and whether FRT regulates LPS-induced cytokine gene expression by modifying post-TLR4 proximal signaling events. Using the murine macrophage RAW 264.7 cell model, we found that exposure to FRT exerted distinct effects on LPS-induced expression of TNF-α, IL-1β, and RANTES without altering global NF-κB or MAPK activation. However, while global NF-κB activation was not affected by exposure to FRT, chromatin immunoprecipitation (ChIP) assays revealed that FRT exposure caused variable effects on recruitment and/or promoter occupancy of HSF1 and NF-κB p65 proteins to the TNF-α and IL-1β gene promoters. These studies suggest that exposure to FRT alters the pattern of cytokine gene expression predominantly through intranuclear events that may modify the recruitment of transcription factors to certain gene promoters in a gene-specific manner.

MATERIALS AND METHODS

HSF1-null mice.

HSF1^−/− mice were derived from a colony established at the University of Maryland, Baltimore, from a breeding pair provided by I. J. Benjamin (29). Homozygous HSF^−/− males were bred with heterozygous females, and genotypes were determined from PCR analysis of tail snip DNA. Mice were housed and all experiments were carried out at the Baltimore Department of Veterans Affairs (VA) Medical Center animal facility under American Association of Laboratory Animal Care-approved conditions and under the supervision of a full-time veterinarian. All protocols were approved by the Institutional Animal Care and Use Committee of the University of Maryland, Baltimore and the Animal Use Subcommittee of the Baltimore VA Medical Center.

Cell culture.

The RAW 264.7 cell line, obtained from the American Type Culture Collection (Manassas, VA), was maintained in RPMI 1640 supplemented with 50 U/ml penicillin, 50 μg/ml streptomycin, 2 mM l-glutamine, 1 mM sodium pyruvate, 10 mM HEPES buffer, pH 7.3 (Invitrogen, Carlsbad, CA), and 10% defined fetal bovine serum (Hyclone, Logan, UT) as described previously (35, 37). Cells were routinely tested for Mycoplasma infection with a commercial assay system (MycoTest, Invitrogen), and new cultures were established monthly from frozen stocks. Cell viability was determined by Trypan blue dye exclusion. Mouse bone marrow-derived macrophages (BMDM) were isolated from femurs of 6- to 10-wk-old HSF1^−/− and heterozygous mice. After euthanasia of mice by isoflurane anesthesia and cervical dislocation, femurs were dissected free and bone marrow cells were aseptically flushed from the marrow cavities with 4°C PBS. Cells were collected by centrifugation, erythrocytes were lysed by resuspending in 0.15 M NH₄Cl for 3–5 min, and cells were washed with PBS and resuspended in complete RPMI medium supplemented with 30% L-929 cell-conditioned medium [as a source of colony-stimulating factor 1 (CSF-1)] and plated. After 7–10 days in culture, adherent macrophages were washed with PBS, scraped gently, counted, and plated in 2 ml at 0.5 × 10⁶/ml in 35-mm dishes.

For FRT/HS exposures, RAW 264.7 cells or BMDM were prewarmed at the target temperature (39.5°C for FRT or 43°C for HS) for 30 min in automatic CO₂ incubators certified to have <0.2°C temperature variation (Forma; Marietta, OH) and calibrated for each experiment with an electronic thermometer (Fluke Instruments model 5211, Everett, WA). After 30-min preincubation, cells were stimulated with either LPS (Escherichia coli 0111B4; Sigma-Aldrich; St. Louis, MO) or N-α-palmitoyl-S-[2,3-bis(palmitoyloxy)-(2RS)-propyl]-l-cysteine (Pam3Cys; Invivogen, San Diego, CA) and maintained at either 39.5°C (FRT) or 37°C (normothermic control). For HS, cells were prewarmed at 43°C for 30 min, stimulated with the agonists, and subsequently incubated at 37°C.

Temperature clamping and intratracheal administration of LPS.

Mice were adapted to standard plastic cages for at least 4 days before the experiment. To avoid the influence of diurnal cycling, all experiments were started at approximately the same time each day (between 8:00 and 10:00 AM). Mice were placed in either 24°C (normothermia) or 34°C (FRT) infant incubators 2 h before LPS instillation, exposures that we previously showed (29) will maintain core temperatures at ∼37°C and 39.5°C, respectively. Six hours after instillation of LPS (50 μg in 50 μl PBS), the mice were euthanized by isoflurane inhalation and cervical dislocation and lung lavage was performed with a total of 2 ml of PBS as previously described (29). TNF-α and IL-1β levels in cell-free supernatants and lung lavage were measured in the University of Maryland Cytokine Core Laboratory (www.cytokines.com) by ELISA using commercial antibody pairs and recombinant standards (R & D Systems).

RNA extraction and quantitative real-time PCR.

Total RNA from 2 × 10⁶ RAW 264.7 cells was isolated with TRIzol (Invitrogen), and contaminating DNA was eliminated with DNase I digestion (Invitrogen). RNA was reverse transcribed with 1 μg of total RNA and oligo(dT) primers with a cDNA synthesis kit according to the manufacturer's protocol (Promega), and real-time PCR reactions were performed as described previously (34). Duplicate 25-μl real-time PCR reactions were performed in 96-well plates with a SYBR Green reaction mix (ABI) and an ABI 7900HT Thermocycler according to the supplier's protocol with the following forward and reverse primers: glyceraldehyde-3-phosphate dehydrogenase (GAPDH), 5′-agcctcgtcccgtagacaaaat and 5′-tggcaacaatctccactttgc; IL-1β, 5′-aaatacctgtggccttgggc and 5′-cttgggatccacactctccag; TNF-α, 5′-gaccctcacactcagatcatcttct and 5′-ccacttggtggtttgctacga; RANTES, 5′-gagtgacaaacacgactgcaagat and 5′-ctgctttgcctacctctccct; and IFN-β, 5′-cacttgaagagctattactggaggg and 5′-ctcggaccaccatccagg. Data were quantified with the gene expression C_t difference method described by Schefe et al. (30) and standardized to levels of the housekeeping gene, GAPDH, with threshold cycle (C_t) values automatically determined by the thermocycler. The efficiency of amplification for each primer pair was calculated, and the expression of the gene of interest in each sample relative to a reference, untreated 37°C cells, was calculated as we previously described (42).

Electrophoretic mobility shift assay.

Nuclear extracts were prepared according to the method of Schreiber et al. (31) as described previously (35, 37), and total protein concentration was measured with the Bradford method (Bio-Rad, Mountain View, CA). Double-stranded oligonucleotides containing the consensus NF-κB binding sequence (Promega, Madison, WI) were radiolabeled with [γ-³²P]ATP using T4 polynucleotide kinase (Promega) according to the manufacturer's protocol. Twenty-microliter EMSA reactions containing 5 μg of nuclear extract, 0.035 pmol of radiolabeled oligonucleotide, 1 μg of poly(dI-dC), 10 mM Tris·HCl pH 7.8, 10% glycerol, 60 mM NaCl, 1 mM EDTA, and 1 mM dithiothreitol were incubated at room temperature for 30 min. The DNA · protein complexes were electrophoretically resolved on 4% nondenaturing polyacrylamide gels. The dried gels were analyzed by phosphorimaging (PhosphorImager, Molecular Dynamics) and subsequently exposed to X-ray film.

Transfection and reporter gene analysis.

NF-κB-luciferase reporter plasmid-containing stable transfectants were generated by transfecting RAW 264.7 cells with pNFκB-luc (Clontech, Mountain View, CA) and the blasticidin resistance plasmid, pcDNA6/TR (Invitrogen), with Fugene 6 (Boehringer Mannheim, Indianapolis, IN). Stable transfectants were selected with 4 μg/ml of blasticidin (Invitrogen), cloned by serial dilution, and maintained in medium containing 2 μg/ml blasticidin. For reporter assays, 1 × 10⁵ cells/ml were plated in blasticidin-free medium in 24-well plates. After 48 h, the cells were treated with 100 ng/ml LPS, incubated at 39.5°C or 37°C for 6 h, and lysed, and luciferase activity was measured with a luciferase assay kit (Promega).

Immunoblotting.

Cell extracts were prepared in RIPA buffer containing 10 mM Tris·HCl pH 8.0, 150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS, protease inhibitors (Boehringer Mannheim), 50 mM sodium fluoride, 20 mM β-glycerophosphate, 1 mM sodium vanadate, and phosphatase inhibitor cocktail (Sigma). For immunoblotting, 20 μg of total protein was resolved by SDS-PAGE, electrostatically transferred to polyvinylidene difluoride membrane (Stratagene), and blocked with either 5% nonfat dry milk or 20% membrane blocking solution (Zymed) in 25 mM Tris pH 7.4, 0.5 M NaCl, 0.05% Tween 20 (TTBS). Membranes were probed with phospho-specific antibodies (Santa Cruz) in 20% membrane blocking solution (Zymed) followed by horseradish peroxidase-conjugated secondary antibody (Bio-Rad), and bands were developed with a chemiluminescence detection system (Renaissance; New England Nuclear, Boston, MA) and exposed to X-ray film. The membrane was then stripped by incubating at 50°C for 30 min in stripping buffer containing 62.5 mM Tris · HCl pH 6.8, 2% SDS, and 100 mM β-mercaptoethanol and reprobed for total proteins (Santa Cruz). The bands were quantified with a gel documentation system (Fuji LAS-1000), and the ratio of phosphorylated to total signal intensity was calculated.

ChIP assays.

ChIP assays were performed with a kit from Millipore (Billerica, MA) as described previously (34, 35). In brief, RAW 264.7 cells were treated as indicated, cross-linked with 1% formaldehyde for 10 min, washed with PBS, and collected by centrifugation. The cell pellets were resuspended in SDS lysis buffer and sonicated for five 10-s bursts with a Branson Sonifier 450 (duty cycle and output settings were 30 and 3, respectively). Sonicated cell lysates were diluted 10-fold with ChIP dilution buffer and precleared for 1 h at 4°C with 80 μl of 50% salmon sperm DNA-saturated protein A agarose beads (ss-protein A). Cross-linked chromatin was immunoprecipitated with 4 μg of primary antibody at 4°C overnight, and immune complexes were collected with ss-protein A. The immune complex was washed and eluted, and cross-linked protein-DNA was reverted by incubating at 65°C for 4 h. DNA was extracted and used as template for real-time PCR as described above. For NF-κB p65 ChIP, rabbit polyclonal anti p65 antibody (Santa Cruz) and primer pairs spanning the NF-κB-binding region on the IL-1β promoter from −338 to −246 (forward 5′-ccgcacatcctgacttaaaatgta and reverse 5′-ttatttccccctggacaattgt) and the TNF-α promoter from −586 to −468 (forward 5′-atgcacacttcccaactctaag and reverse 5′-cttctgaaagctgggtgcataag) were used. For HSF1 ChIP, HSF1 rabbit polyclonal antibody (Santa Cruz) and primer pairs spanning the putative HSF1-binding region from the IL-1β promoter (forward 5′-cccctaagaattcccatcaagc and reverse 5′-accactgcagggtttgttgtc) and the TNF-α promoter (forward 5′-agcgaggacagcaaggga and reverse 5′-tcttttctggagggagtgtgg) were used. For acetyl-histone H3 ChIP, anti-acetyl histone H3 rabbit polyclonal antibody (no. 06-599, Millipore) was used and real-time quantitative RT-PCR (qRT-PCR) was performed with the primers spanning the NF-κB-binding region on the IL-1β and TNF-α promoters. ChIP qRT-PCR data were analyzed with a template from SA Biosciences, and fold enrichment with specific antibody against nonspecific rabbit polyclonal antibody was determined.

Statistics.

Data are displayed as means ± SE. Differences between two groups were analyzed by unpaired Student t-test, and differences among multiple groups were analyzed by applying a Tukey-Kramer honestly significant difference test to a one-way ANOVA. Differences between groups were accepted at P < 0.05.

RESULTS

FRT exposure differentially modifies LPS-induced cytokine gene expression activated through a common proximal signaling pathway.

While we previously showed (35) that FRT exposure represses TNF-α transcription, in part by activating HSF1 to bind to an HSE in the 5′-untranslated region (UTR) of the murine TNF-α gene, the results of that study also suggested that FRT exposure represses TNF-α transcription through additional mechanisms. TLR4 engagement by LPS stimulates bifurcating MyD88-dependent and -independent signaling pathways (5, 21, 43). To determine whether FRT reduces LPS-induced TNF-α expression by altering proximal signaling pathways, we analyzed the effect of FRT exposure on LPS-induced expression of TNF-α (Fig. 1A) and IL-1β (Fig. 1B), two genes activated via the same MyD88-dependent proximal signaling pathway, and IFN-β (Fig. 1C) and RANTES (Fig. 1D), two genes activated via a common MyD88-independent pathway (11, 41). For FRT exposure, RAW 264.7 cells were prewarmed at 39.5°C for 30 min, stimulated with 100 ng/ml LPS, and incubated at 39.5°C as we described in our earlier studies (35–37). Normothermic controls were stimulated with LPS and maintained at 37°C. Treatment with LPS increased TNF-α mRNA levels in both the 37°C and 39.5°C RAW 264.7 cells, but the TNF-α mRNA levels 1 and 3 h after LPS stimulation were 57% and 44% lower in cells incubated in 39.5°C cultures (Fig. 1A), consistent with our previously reported findings (37). In contrast, IL-1β mRNA, which is activated by LPS via the same MyD88-dependent pathway, exhibited an opposite response to FRT, increasing by 2.6-fold and 1.7-fold at 1 and 3 h after LPS in 39.5°C cell culture, respectively, compared with cells stimulated with LPS at 37°C (Fig. 1B). Two cytokines activated by LPS through the MyD88-independent signaling pathway were also differentially modulated by FRT. While IFN-β mRNA levels were similar in 37°C and 39.5°C cells (Fig. 1C), RANTES mRNA levels were reduced by 37% in the FRT-exposed cells (Fig. 1D). That FRT exerts dissimilar effects on cytokines that share common proximal signaling pathways suggests that these effects are not mediated exclusively through a single alteration in proximal signaling.

Fig. 1.

LPS-induced cytokine expression profile is altered by exposure to febrile-range temperature (FRT). RAW 264.7 macrophages were prewarmed to 37°C or 39.5°C for 30 min and then stimulated with 100 ng/ml LPS; incubation continued at the same temperature for an additional 1 or 3 h, and cells were lysed for collection of RNA. Cytokine mRNA was measured by quantitative real-time RT-PCR, expressed as a ratio to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA levels, and presented as fold change compared with the time 0 measurement. A: TNF-α levels. B: IL-1β levels. C: IFN-β levels. D: regulated upon activation, T cell expressed and secreted (RANTES) levels. Data are means ± SE of 6 independent experiments. Data from unstimulated and LPS-stimulated cells are shown at left and right. *P < 0.05 compared with time 0; †P < 0.05 vs. 37°C cells at the same time point.

Heat shock exposure modifies LPS-induced cytokine expression in a manner distinct from FRT exposure.

Because temperatures in the febrile and HS ranges may exert distinct effects on cellular processes (16) and because other groups (1, 45) have reported that HS represses, rather than augments, LPS-induced IL-1β expression in human monocyte-like THP-1 cells, we compared the effect of HS and FRT on expression of TNF-α, IL-1β, IFN-β, and RANTES (Fig. 2). HS was induced in RAW 264.7 cells by 30-min exposure to 43°C. The cells were then treated with 100 ng/ml LPS and transferred to 37°C. HS and FRT caused similar reductions in LPS-induced TNF-α expression (Figs. 1A, 2A) but exerted distinct effects on the other three cytokines studied. IL-1β expression, which was enhanced in FRT-exposed cells, was mildly repressed in heat-shocked cells (Fig. 2B). Expression of IFN-β, which was not substantially altered by exposure to FRT, was reduced by 42% in HS cells (Fig. 2C). RANTES expression, which was reduced in FRT-exposed cells, was not affected by exposure to HS (Fig. 2D). These results indicate that FRT and HS exposure might regulate the expression of cytokines through distinct mechanisms.

Fig. 2.

LPS-induced cytokine expression profile is differentially altered by exposure to heat shock (HS). RAW 264.7 macrophages were prewarmed to 37°C or 43°C for 30 min and then stimulated with 100 ng/ml LPS; incubation continued for an additional 1 or 3 h at 37°C, and cells were lysed for collection of RNA. Cytokine mRNA was measured by quantitative real-time RT-PCR, expressed as a ratio to GAPDH mRNA levels, and presented as fold change compared with time 0 ratios. A: TNF-α levels. B: IL-1β levels. C: IFN-β levels. D: RANTES levels. Data are means ± SE of 6 independent experiments. *P < 0.05 compared with time 0; †P < 0.05 vs. 37°C cells at the same time point.

FRT exposure represses TNF-α and enhances IL-1β expression stimulated by TLR2 agonist Pam3Cys.

To further analyze the effects of FRT on MyD88-dependent expression of TNF-α and IL-1β genes, we analyzed cytokine gene expression in cells stimulated with the TLR2 agonist Pam3Cys (Fig. 3). Since TLR2 shares the MyD88 signaling pathway with TLR4 but does not activate the MyD88-independent pathway (41), we restricted the analysis to TNF-α and IL-1β. The effect of FRT on TNF-α and IL-1β mRNA expression was similar in Pam3Cys- and LPS-stimulated cells. Pam3Cys-induced TNF-α mRNA levels were reduced by 83% in FRT-exposed cells, while IL-1β mRNA levels were increased by 2.4-fold compared with 37°C cells (49).

Fig. 3.

Pam3Cys-induced cytokine expression profile is altered by exposure to FRT. RAW 264.7 macrophages were prewarmed to 37°C or 39.5°C for 30 min and then stimulated with 1 μg/ml Pam3Cys; incubation continued for an additional 1 or 3 h at the same temperature, and cells were lysed for isolation of RNA. TNF-α (A) and IL-1β (B) mRNA was measured by quantitative real-time RT-PCR, expressed as a ratio to GAPDH mRNA levels, and presented as fold change compared with the time 0 measurement. Data are means ± SE of 6 independent experiments. Data from unstimulated and LPS-stimulated cells are shown at left and right, respectively. *P < 0.05 compared with time 0; †P < 0.05 vs. 37°C cells at the same time point.

Kinetics of FRT exposure on RNA levels of TNF-α and IL-1β in LPS-stimulated RAW 264.7 cells.

Since exposure to FRT unexpectedly enhanced rather than suppressed IL-1β expression, we performed a more complete analysis of the kinetics of LPS-induced TNF-α and IL-1β expression. RAW 264.7 cells were treated with LPS at 37°C and 39.5°C as described above, but we analyzed mRNA levels every 30 min between 1 h and 4 h after LPS stimulation (Fig. 4). As expected, TNF-α mRNA levels (Fig. 4A) increased in both 37°C- and 39.5°C-exposed cells after LPS stimulation and the increase was suppressed in the 39.5°C cells. In contrast, LPS-induced IL-1β mRNA levels (Fig. 4B) were increased in the 39.5°C cells after LPS stimulation and persisted for the 4-h poststimulation period.

Fig. 4.

Kinetics of LPS-induced TNF-α and IL-1β mRNA expression in RAW 264.7 cells incubated at 37°C and 39.5°C. RAW 264.7 macrophages were prewarmed to 37°C or 39.5°C for 30 min and then stimulated with 100 ng/ml LPS; incubation continued at the same temperature for up to 4 h. Aliquots of cells were lysed every 30 min for collection of RNA. TNF-α (A) and IL-1β (B) mRNA was measured by quantitative real-time RT-PCR, expressed as a ratio to GAPDH mRNA levels, and presented as fold change compared with the time 0 measurement. Data are means ± SE of 3 independent experiments. *P < 0.05 vs. 37°C cells at the same time point.

FRT-induced repression of TNF-α is HSF1 dependent.

We previously showed (35) that exposure to FRT activates HSF1:HSE binding activity and represses TNF-α transcription in the RAW 264.7 mouse macrophage cell line, suggesting a role for HSF1 in FRT-induced TNF-α repression. To determine whether HSF1 is required for FRT-induced TNF-α repression, we measured LPS-induced TNF-α secretion at 37°C and 39.5°C in BMDM isolated from HSF1^−/− mice and their HSF1^+/− littermates (Fig. 5A). Cells from HSF1^+/− mice secreted ∼40% less TNF-α when stimulated with 100 ng/ml LPS at 39.5°C compared with cells stimulated with LPS at 37°C. In contrast, a similar warming protocol did not reduce TNF-α secretion in HSF1^−/− macrophages. We could not detect any IL-1β secretion in the BMDM in the presence or absence of LPS and FRT (data not shown). We also analyzed in vivo intrapulmonary TNF-α and IL-1β expression in HSF1^−/− and HSF1^+/− mice challenged intratracheally with 50 μg of LPS (Fig. 5, B and C). After LPS instillation, mice were maintained for 6 h at either normothermic core temperature (37°C) or FRT (core temperature ∼39.5°C) and TNF-α and IL-1β levels were measured by ELISA in lung lavage. Exposure to FRT reduced TNF-α concentrations in lung lavage by ∼40% in the heterozygous mice but failed to modify TNF-α levels in HSF1-null mice. In contrast, LPS-induced IL-1β level was ∼25% greater in the FRT-exposed than normothermic groups in both HSF1^−/− and HSF1^+/− mice. Collectively, these experiments demonstrated that HSF1 is required for repression of LPS-induced TNF-α expression but not enhancement of IL-1β expression at FRT.

Fig. 5.

TNF-α secretion is reduced at FRT in heat shock factor-1 (HSF1)^+/− but not HSF1^−/− macrophages and mice. A: bone marrow-derived macrophages from HSF1^−/− and HSF1^+/− mice were prewarmed to 37°C or 39.5°C for 30 min and then stimulated with 100 ng/ml LPS; incubation continued for an additional 6 h, and supernatants were collected and analyzed for TNF-α concentration by ELISA. Data are means ± SE of 3 separate experiments. *P < 0.05 compared with 37°C for each genotype. B and C: HSF1^−/− and HSF1^+/− mice were housed at 24°C (euthermic) or 34°C (FRH) ambient temperature for 2 h until core temperature equilibrated. Each mouse then received 50 μg of LPS via intratracheal instillation and, 6 h later, was euthanized, and lung lavage (bronchoalveolar lavage fluid, BALF) was analyzed for TNF-α (B) or IL-1β (C) concentration by ELISA. Data are means ± SE of 5 mice/group. *P < 0.05 vs. euthermic for each genotype.

LPS-induced activation of ERK and p38 MAPK and activation of NF-κB is unaffected by exposure to hyperthermia.

Since FRT exerts opposing effects on TNF-α and IL-1β expression, one cannot attribute all effects to a single modification in their common proximal MyD88-dependent signaling pathway. Nonetheless, some of the effects of FRT on cytokine expression may result from one or more alterations in TLR4-activated signaling pathways. To evaluate this possibility, we analyzed the effects of FRT exposure on activation of ERK and p38 MAPK and activation of NF-κB, three important signaling intermediates that are activated by LPS and participate in activation of TNF-α and IL-1β gene expression (5). Activation of ERK and p38 MAPK, as measured by Western blotting for the phosphorylated form of each kinase (Fig. 6, A–D) or their downstream substrate kinases ribosomal S6 kinase (RSK) and MAPKAPK2, respectively (data not shown), revealed similar activation kinetics after LPS stimulation in both 37°C and 39.5°C cell cultures. LPS-induced activation of NF-κB as measured by EMSA (Fig. 6, E and F) or the activity of a synthetic NF-κB-responsive reporter construct (pNF-κB, Clontech) in stably transfected RAW 264.7 cells (Fig. 6G) showed no appreciable difference in LPS-stimulated RAW 264.7 cells in 37°C and 39.5°C cell cultures. Likewise, exposure to HS at 43°C for up to 60 min also exerted no appreciable effect on LPS-induced activation of NF-κB activation. Longer exposures to 43°C resulted in reduced ERK, p38, and NF-κB activation that was associated with extensive cell death. These data clearly demonstrate that the effects of FRT exposure on LPS-induced TNF-α or IL-1β expression occur without detectable alteration in global activation of three important post-TLR4 receptor signaling intermediates, ERK, p38 MAPK, and NF-κB.

Fig. 6.

Exposure to FRT fails to alter LPS-induced activation of NF-κB, ERK, and p38 MAPK. A–D: RAW 264.7 macrophages were prewarmed to 37°C, 39.5°C, or 43°C (if indicated) for 30 min and then stimulated with 100 ng/ml LPS; incubation continued at the same temperature for the indicated time, cell lysates were analyzed for levels of phosphorylated and total ERK (A, B) and p38 (C, D) by Western blot, and band density was quantified by direct imaging of chemiluminescent signal. A representative blot (A, C) and mean ± SE of densitometric values from 4 separate experiments (B, D) are shown. E and F: RAW 264.7 cells were treated as in A–D, and nuclear lysates were collected and in vitro NF-κB-DNA binding measured by EMSA. A representative autoradiograph (E) is shown as well as the means ± SE of EMSA band density determined by phosphorimaging from 4 separate experiments (F). G: RAW 264.7 cells stably transfected with an NF-κB-responsive luciferase reporter construct were prewarmed to 37°C, 39.5°C, or 43°C for 30 min and then stimulated with 100 ng/ml LPS; incubation continued at the same temperature for 6 h, and luciferase activity in cell lysates was measured. Means ± SE of 4 separate experiments are shown. *P < 0.05 compared with 43°C/15-min exposure.

FRT exposure differentially regulates recruitment of NF-κB p65 to TNF-α and IL-1β gene promoters.

Since FRT exposure showed no significant effect on the global activation of either MAPKs or NF-κB, we speculated that the gene-specific effects of FRT might be achieved by modifying interactions between one or more transcription factors and the gene promoters. Since NF-κB is the common regulator of both TNF-α and IL-1β transcription (14, 32), we analyzed the effect of FRT on recruitment of NF-κB p65 to the TNF-α and IL-1β gene promoters. RAW 264.7 cells were stimulated with LPS for 30 min at 37°C or prewarmed at 39.5°C for 30 min and stimulated with LPS for 30 min at 39.5°C. Recruitment of NF-κB p65 to the promoters of each cytokine gene was analyzed by ChIP assay, and immunoprecipitated chromatin was quantified by qRT-PCR. As expected, LPS markedly enhanced NF-κB p65 recruitment to both TNF-α (Fig. 7A) and IL-1β (Fig. 7B) promoters in 37°C-treated cells. However, exposure of cells to 39.5°C caused opposing effects of NF-κB p65 recruitment to the two promoters, reducing recruitment in the TNF-α promoter by ∼40% while simultaneously increasing recruitment to the IL-1β promoter by ∼60% compared with LPS-stimulated cells maintained at 37°C. To determine whether the effect of FRT on NF-κB p65 recruitment to the TNF-α or IL-1β promoter may be caused by temperature-dependent alterations in chromatin modification, we analyzed histone H3 acetylation of the p65 binding regions of the two genes by ChIP using anti-acetyl-histone H3 antibody (Fig. 7, C and D). As expected, LPS significantly enhanced histone H3 acetylation at the target TNF-α and IL-1β promoter regions in cells at 37°C and coexposure to 39.5°C abrogated the same at the TNF-α promoter. Surprisingly, rather than enhancing histone H3 acetylation, exposure of the cells to 39.5°C reduced histone H3 acetylation at the NF-κB-binding region of IL-1β promoter, albeit to a lesser extent than TNF-α gene promoter.

Fig. 7.

FRT differentially modifies recruitment of NF-κB-p65 to TNF-α and IL-1β promoters. A and B: RAW 264.7 macrophages were prewarmed to 37°C or 39.5°C for 30 min, stimulated with 100 ng/ml LPS, and incubated at the same temperature for another 30 min. Cross-linked chromatin was analyzed for recruitment of NF-κB p65 to the TNF-α (A) and IL-1β (B) promoters by chromatin immunoprecipitation (ChIP) assay and expressed as fold change vs. unstimulated 37°C cells. C and D: chromatin from the same cells was also analyzed by ChIP for acetylated histone H3 at the same regions of the TNF-α (C) and IL-1β (D) promoters. Data are means ± SE of 4 independent experiments. *P < 0.05 compared with untreated controls; †P < 0.05 vs. 37°C cells at the same time point.

Earlier studies by our group (9) as well as by others showed (1) that exposure to FRT and HS repressed IL-1β expression in human monocytes and macrophages. Xie et al. (45) demonstrated that HSF1, activated by exposure to HS, repressed human IL-1β expression by interfering with the trans-activating activity of C/EBPβ. This effect required intact binding elements for both HSF1 and C/EBPβ, which are present in close proximity in the human IL-1β promoter (45). However, analysis of the same region of the mouse IL-1β promoter revealed a possible disruption of the canonical HSF1 binding sequence by interposition of an additional adenine nucleotide between the inverted dyad nGAAn repeats (Fig. 8A), which may explain the lack of repression of murine IL-1β by FRT/HSF1. To evaluate the functional consequences of the altered HSE sequence in the mouse IL-1β promoter, we compared recruitment of HSF1 to the murine TNF-α 5′-UTR (Fig. 8B) and the proximal murine IL-1β promoter (Fig. 8C) in mouse RAW 264.7 cells exposed to 39.5°C for 1.5–2.5 h. While exposure to FRT increased recruitment of HSF1 to the TNF-α 5′-UTR, as we have previously demonstrated (35), there was no detectable effect of such exposure on HSF1 recruitment to the IL-1β promoter, indicating the absence of any functional interaction between HSF1 and the proximal mouse IL-1β promoter.

Fig. 8.

HSF1 is differentially recruited to the proximal murine IL-1β and TNF-α promoters. A: sequences of the heat shock response element (HSE)-containing human IL-1β promoter sequence and the analogous mouse IL-1β sequence are compared. C/EBPβ-binding sequence and HSE are underlined and bold. The extra adenine in the mouse sequence is shown in parentheses. B and C: RAW 264.7 cells were incubated for the indicated time at either 37°C or 39.5°C and analyzed for HSF1 recruitment to the HSE-containing sequences of the TNF-α 5′-UTR (B) and the IL-1β promoter (C) by ChIP assay. Data are means ± SE of 4 separate experiments. *P < 0.05 compared with time 0.

DISCUSSION

In earlier studies we demonstrated that FRT exposure reduced TNF-α transcription in human and mouse macrophages (8, 9, 35–37), an effect that was mediated through HSF1 interaction with the TNF-α promoter (35, 37). The present study extends these observations to show that FRT augments, rather than represses, IL-1β expression and that the opposing effects of FRT on TNF-α and IL-1β expression occur without detectable changes in TLR4 agonist-induced global activation of three important signaling intermediates, NF-κB and ERK and p38 MAPKs. Interestingly, while activation and nuclear translocation of NF-κB were unaffected by FRT, recruitment of NF-κB to the TNF-α and IL-1β promoters was differentially modified by FRT and paralleled the effects on gene transcription.

TNF-α secretion by BMDM stimulated in vitro with LPS and intrapulmonary TNF-α and IL-1β accumulation stimulated by intratracheal LPS instillation demonstrated the same pattern of temperature- and HSF1 dependence. Since Xiao et al. (44) showed that heterozygous HSF1 mice possessing a single intact HSF1 allele exhibit wild-type levels of HS responsiveness, we compared HSF1-null mice and their heterozygous littermates and found that FRT repressed LPS-induced TNF-α expression by ∼40% in HSF1^+/− but did not alter TNF-α expression in HSF1^−/− mice. In contrast, IL-1β levels were augmented in both HSF1^+/− and HSF1^−/− mice.

We previously showed (35) that exposure to FRT may repress TNF-α expression through multiple mechanisms, including one that does not require HSF1 binding to the HSE in the TNF-α 5′-UTR. Likewise, Chen et al. (4) demonstrated that an HSF1 mutant that was defective in HSE binding capacity was still capable of repressing c-fos transcription in CHO K1 cells. HSF1 has been shown to interact with several transcription factors, including signal transducer and activator of transcription-1 (STAT-1), C/EBPβ, TATA-binding protein (TBP), and transcription factor complexes TFIIB and D (38, 45, 46, 49), with variable consequences for transcription. While the opposing effects of FRT on TNF-α and IL-1β are unlikely to be caused by a single alteration in post-TLR signaling events, this did not exclude potentially important effects of FRT on post-TLR signaling pathways. To evaluate this possibility further, we analyzed three post-TLR signaling events, namely, activation of NF-κB and ERK and p38 MAPK in RAW 264.7 cells stimulated with LPS at 37°C and 39.5°C, and found no difference in the magnitude or kinetics of activation.

Other groups have reported that exposure of cells to hyperthermia reduces NF-κB activation, but most of these studies focused on temperatures in the HS range (22, 23, 26, 27). Malhotra et al. (26) showed that HSF1 is not required for inhibition of TNF-α-induced NF-κB activation in fibroblasts exposed to 43°C for 1 h. Lee et al. (22, 23) and Pittet et al. (27) showed that exposure to 43°C for at least 1 h inactivates IKKα and IKKβ, NF-κB-inducing kinase (NIK), and IL-1 receptor associated-kinase (IRAK)4, an effect that is reversed by heat shock protein (HSP)72 and HSP90. In contrast, Shi et al. (33) showed that overexpression of HSP72 reduced LPS-induced NF-κB p65 translocation in RAW 264.7 cells. Ran and coworkers (28) found that overexpression of HSP72 inhibited TNF-α-induced activation of an NF-κB reporter plasmid and showed that HSP72 bound to IKKγ and blocked recruitment of IKKα and IKKβ to the IKK complex. In contrast to the findings of the current study, Hagiwara et al. (15) found that LPS-induced NF-κB activation was reduced in RAW 264.7 cells exposed to 40°C. However, in addition to the slightly higher exposure temperature used in the Hagiwara study, these investigators only analyzed NF-κB activation 3 h after stimulation with LPS, when much of the TNF-α and IL-1β transcription has already occurred, as shown in Figs. 1 and 4 of this study and demonstrated in our earlier studies by nuclear run on assays (37).

Several studies have shown that despite comparable nuclear translocation and DNA binding activity of NF-κB, differences in chromatin modification may differentially modify LPS-induced transcription of various genes (2, 7, 47, 48). This led us to speculate that FRT exposure might differentially modify recruitment of NF-κB proteins to TNF-α and IL-1β gene promoters. In fact, ChIP assays using antibody against NF-κB p65 protein showed that FRT exposure enhanced LPS-induced NF-κB p65 recruitment to the IL-1β promoter, indicating that this enhanced occupancy of NF-κB at the IL-1β chromatin might account for the enhanced expression of the gene in FRT-exposed cells. Furthermore, the reduced recruitment of NF-κB to the TNF-α promoter may represent a second mechanism of FRT-induced repression of TNF-α gene expression, in addition to direct binding of HSF1 to the TNF-α 5′-UTR.

We reasoned that the altered recruitment of NF-κB to the TNF-α and IL-1β promoters in FRT-exposed cells might result from disparate effects of FRT on chromatin architecture of the two gene promoters. However, a ChIP analysis for acetylated histone H3 demonstrated reduced LPS-induced acetylation of both promoters at 39.5°C. While temperature-dependent effects on histone H3 acetylation may explain the observed reduction in NF-κB recruitment to the TNF-α promoter at 39.5°C, they do not explain the concurrent increase in IL-1β gene activation. In an earlier study, Liang et al. (25) demonstrated that the IL-1β gene is inducibly transcribed from a poised promoter architecture and that the IL-1β chromatin is equivalently accessible to transcription factors in resting and stimulated states, implicating that the recruitment of the transcription factor rather than accessibility was the rate-limiting factor.

Hagiwara et al. (15) reported that RAW 264.7 cells secreted less IL-1β when activated with LPS at 40°C rather than 37°C. However, there are complexities associated with IL-1β maturation and secretion that modify IL-1β secretion in vitro (10, 12). Hence in the present study, we measured IL-1β mRNA levels rather than IL-1β protein secretion and found that LPS-induced IL-1β mRNA was enhanced after incubation at 39.5°C. Our finding that RAW 264.7 mouse macrophages increase IL-1β expression at FRT appears to contradict findings by Cahill and Xie and coworkers (1, 45), who showed that HS repressed IL-1β expression in human monocytes and macrophages by interfering with the trans-activating activity of C/EBPβ. However, this effect required intact binding elements for both HSF1 and C/EBPβ, which are present in close proximity in the human IL-1β promoter (37). As we have shown (Fig. 7A), the mouse IL-1β promoter contains an additional adenine that disrupts the consensus HSE adjacent to the C/EBPβ site, and our ChIP assay confirmed that the HSF1-binding function has been disrupted. Thus despite HSF1 activation at FRT (37), HSF1 fails to exert any repressive effect on murine IL-1β gene transcription at FRT. Moreover, Cahill et al. heat shocked cells by exposing them to 43°C compared with 39.5°C in our FRT exposure study. When we exposed RAW 264.7 cells to 43°C, we also found a decrease in IL-1β message level compared with 37°C cells (Fig. 2), indicating that exposure to HS and FRT exert different effects on activation of cytokine genes, including IL-1β.

The fact that hyperthermic exposures and HSF1 could modify chromatin architecture has been reported previously (6, 17, 20, 40). In a recent report, Inouye et al. (17) showed that HSF1 facilitated the recruitment of LPS-induced NF-κB to the IL-6 promoter. These results, coupled with our earlier reports (35–37) that HSF1 mediates the repression of TNF-α expression during exposure to FRT, indicate that FRT modifies cytokine gene expression in a gene-specific manner, rather than by modifying common proximal signaling events. Furthermore, these results suggest that FRT-induced modification of gene expression may occur, in part, through altered, gene-specific, recruitment of transcription factors to gene promoters.

In summary, we have found that exposure to hyperthermia at levels that reflect temperatures achieved during fever and clinically relevant heat-related illnesses differentially regulates cytokine gene expression. These results demonstrate that exposure to such temperatures may exert important pleiotropic effects that could modify the innate immune response during febrile illnesses and exertional and environmental hyperthermia.

GRANTS

This work was supported by National Institutes of Health Grants GM-069431 (I. S. Singh), GM-066855, HL-69057, and HL-085256 (J. D. Hasday), and AI-18797 (S. N. Vogel) and by VA Merit Review grants to J. D. Hasday and I. S. Singh.

DISCLOSURES

No conflicts of interest are declared by the author(s).