Febrile-Range Hyperthermia Augments Neutrophil Accumulation and Enhances Lung Injury in Experimental Gram-Negative Bacterial Pneumonia
http://www.100md.com
免疫学杂志 2005年第6期
Abstract
We previously demonstrated that exposure to febrile-range hyperthermia (FRH) accelerates pathogen clearance and increases survival in murine experimental Klebsiella pneumoniae peritonitis. However, FRH accelerates lethal lung injury in a mouse model of pulmonary oxygen toxicity, suggesting that the lung may be particularly susceptible to injurious effects of FRH. In the present study, we tested the hypothesis that, in contrast with the salutary effect of FRH in Gram-negative peritonitis, FRH would be detrimental in multilobar Gram-negative pneumonia. Using a conscious, temperature-clamped mouse model and intratracheal inoculation with K. pneumoniae Caroli strain, we showed that FRH tended to reduce survival despite reducing the 3 day-postinoculation pulmonary pathogen burden by 400-fold. We showed that antibiotic treatment rescued the euthermic mice, but did not reduce lethality in the FRH mice. Using an intratracheal bacterial endotoxin LPS challenge model, we found that the reduced survival in FRH-treated mice was accompanied by increased pulmonary vascular endothelial injury, enhanced pulmonary accumulation of neutrophils, increased levels of IL-1, MIP-2/CXCL213, GM-CSF, and KC/CXCL1 in the bronchoalveolar lavage fluid, and bronchiolar epithelial necrosis. These results suggest that FRH enhances innate host defense against infection, in part, by augmenting polymorphonuclear cell delivery to the site of infection. The ultimate effect of FRH is determined by the balance between accelerated pathogen clearance and collateral tissue injury, which is determined, in part, by the site of infection.
Introduction
Fever is generally regarded as a protective response to infection (1). Retrospective clinical studies demonstrate an association between the presence of fever and reduced length of viral illnesses (2, 3, 4, 5) or improved survival in bacterial infections (6, 7, 8, 9, 10). However, this association is lost in patients with higher disease acuity (11). In two retrospective clinical studies of patients with bacteremia and sepsis, treatment with the antipyretic agent, acetaminophen, was associated with improved survival, although actual core temperatures were not mentioned in the report (12, 13). Studies in animal models in which febrile-range hyperthermia (FRH)3 was achieved by increasing ambient temperature generally support the concept that a temperature increase to febrile levels is protective during infections (14, 15, 16, 17, 18, 19, 20, 21, 22). When the desert lizard, Diposaurus dorsalis, is inoculated s.c. with the Gram-negative pathogen, Aeromonas hydrophila, it increases its core temperature by seeking a warmer environment (14, 18, 21). In this model, a 2°C increase in core temperature improves survival, accelerates pathogen clearance, and augments accumulation of granulocytes at the inoculation site (18). Similarly, we have shown that housing mice with experimental Klebsiella pneumoniae peritonitis at 34.5°C rather than 23°C increased core temperature from 37 to 39.7°C, reduced i.p. bacterial load by five orders of magnitude, and increased survival from 0 to 50% (20). In that study, in vitro K. pneumoniae proliferation rates were comparable in 37 and 39.5°C bacterial cultures, indicating that the accelerated bacterial clearance in the warmer animals was mediated through enhanced host defense rather than direct bactericidal effects of FRH. Interestingly, while death in the euthermic mice was associated with overwhelming infection, death in the hyperthermic mice occurred in the face of much lower bacterial burdens. The discordance between lethality and bacterial burden in the warmer mice suggested a greater contribution from the host immune response to lethality in the presence of FRH. A similar discordance between bacterial load and survival was found by Ellingson et al. (23) in rabbits with experimental Streptococcus pneumoniae. Based on these studies, we reasoned that the ultimate effect of fever would be determined by a dynamic balance between accelerated pathogen clearance and augmented collateral host tissue injury, and would depend, in part, on the nature of the organ system involved. Because of its delicate anatomic structure and its critical gas exchange function, the lower respiratory tract might be less tolerant of the immunostimulatory effects of FRH. We recently found that lethal lung inflammation and injury is augmented by FRH in a mouse model of pulmonary oxygen toxicity (24), illustrating the potential harm of FRH-augmented lung injury. However, hyperoxia is an artificial injurious agent that stimulates injury in the absence of a replicating pathogen.
In the present study, we tested the hypothesis that, in contrast with its salutary effect in Gram-negative peritonitis (20), FRH would be detrimental in diffuse Gram-negative pneumonia because the augmented collateral lung injury would offset any enhancement of pathogen elimination. We used a conscious, temperature-clamped mouse model to elucidate the effects of FRH on the outcome of K. pneumoniae pneumonia, antibiotic-treated K. pneumoniae pneumonia, and intratracheal (i.t.) bacterial endotoxin LPS challenge. To further analyze the effects of FRH on collateral tissue injury, we focused on the potential contribution of polymorphonuclear cells (PMNs) and the role of CXC chemokines and GM-CSF in the FRH-augmented lung injury.
Materials and Methods
Mice and reagents
Eight- to 10-wk-old male outbred CD-1 mice, weighing 25–30 g were purchased from Harlan Sprague Dawley, housed in the Baltimore Veterans Administration Medical Center animal facility under American Association of Laboratory Animal Care-approved conditions and under the supervision of a full-time veterinarian. All animals were used within 4 wk of delivery. Heat shock factor-1 (HSF-1)-null mice have been previously characterized (25) and a colony established at the University of Maryland. Homozygous HSF-1-null males were bred with heterozygous females and the genotypes determined from PCR analysis of tail snip DNA as previously described (25). LPS prepared by trichloroacetic acid extraction from Escherichia coli O111:B4 and 2,2,2-tribromoethyl were purchased from Sigma-Aldrich. All protocols were approved by the Insitutional Animal Care and Use Committee of the University of Maryland, Baltimore.
Bacterial culture
K. pneumoniae 1:K2 strain B5055 and the Caroli substrain of B5055 were obtained from Dr. I. Orskov and Dr. F. Orskov (State Serum Institute, Copenhagen, Denmark) and stored as frozen glycerol stocks. The bacteria were cultured in tryptic soy broth (Sigma-Aldrich) at 37°C and growth was monitored by measuring OD650. Bacteria were harvested during logarithmic growth, washed three times with sterile PBS (pH 7.2), adjusted to 0.3 OD units (1 x 108 bacteria per milliliter), and used to inoculate mice. Inoculum size was confirmed by enumerating colony growth on MacKonkey agar.
Temperature clamping, experimental pneumonia, and i.t. administration of LPS
Mice were adapted to standard plastic cages for at least 4 days before study. To avoid the influence of diurnal cycling, all experiments were started at approximately the same time each day (between 8:00 and 10:00 a.m.). To measure core temperature, 1 wk before a planned experiment, one sentinel mouse from each experimental group was anesthetized with 250 mg/kg tribromoethanol (Avertin) i.p. A sterilized Mini Mitter 1.05g in vivo temperature sensor, model 100-0035, was freely placed into the peritoneal cavity, the peritoneum and abdominal musculature were closed with a single layer of sutures, and the skin closed using surgical glue (Vet-Bond). Euthermia was maintained by holding the mice in a 36.5°C infant isolette until they awakened. The mice received one dose of 2.5 mg/kg buprenorphine analgesia s.c. immediately after surgery and food and water were provided ad libitum. The signal from the temperature sensors was continuously monitored using the Mini Mitter Automated Data Acquisition System.
On the day of inoculation, mice were anesthetized with 250 mg/kg Avertin i.p. and placed supine at a 45° angle. The tongues were gently extended using forceps, 50 μl of K. pneumoniae inoculum or 1 mg/ml LPS in sterile PBS was pipetted gently into the posterior pharynx, and the mice held in this position until aspiration was confirmed as described by Keane-Myers et al. (26). In preliminary experiments, we showed that Evans blue dye administered using this method stained the surfaces of all lobes of both lungs with only small amounts swallowed or retained in the trachea. To prevent anesthetic-induced hypothermia, euthermia was maintained in a 36.5°C isolette until the mice awakened. Upon awakening, cages containing groups of two to three mice were transferred to either 24 or 34°C infant isolettes for the remainder of the experiment. In conscious unrestrained mice, such exposures maintain core temperatures of 36.5–37°C and 39.5–40°C, respectively (20). Room temperature was 23–24°C. To study the effect of FRH on survival in antibiotic-treated pneumonia, mice were inoculated i.t with 250 CFU K. pneumoniae and returned to 24°C cages for 48 h. Then half the mice were switched to 34.5°C isolettes (FRH) and treatment was initiated with 28 mg/kg ceftriaxone i.p per day.
Lung lavage and processing
At selected times, groups of mice were anesthetized by 10–30 s exposure to isoflurane, euthanized by cervical dislocation, and lung lavage was performed in situ through an 18-gauge blunt-end needle secured in the trachea, using 1 ml of PBS instilled and withdrawn twice, followed by instillation and recovery of a second 1 ml of PBS. The two aliquots of lung lavage were pooled, cells were collected by centrifugation at 1000 x g for 3 min, and cell-free supernatants were stored at –80°C for analysis of total protein and cytokine concentrations. Total cell counts were performed manually using a hemacytometer and differential cell counts of DiffQuick-stained cytopreparations (Baxter Scientific Products) were performed by two blinded observers (P. Rice and J. D. Hasday) using morphologic criteria.
Measurement of bacterial load
The lungs were weighed and homogenized in 1 ml of 0.9% NaCl between sterile frosted glass slides. The solid tissue was allowed to sediment for 10 min at room temperature. Homogenate supernatants and blood were serially diluted and 10-μl aliquots of each were plated on MacKonkey agar and incubated at 37°C overnight, and the number of colonies was enumerated and expressed per gram wet weight of lung tissue.
Measurement of cytokine and total protein concentration
Mouse TNF-, IFN-, IL-1, GM-CSF, MIP-2, and KC and human IL-8 were measured in the University of Maryland Cytokine Core Laboratory using standard two-Ab ELISA with commercial Ab pairs and recombinant standards (TNF- and IL-6 from Endogen; IFN-, GM-CSF, MIP-2, and KC from R&D Systems; and IL-1 from Genzyme) and IL-8 from BioSource International. Polystyrene plates (Maxisorb; Nunc) were coated with capture Ab in PBS overnight at 25°C. The plates were washed four times with 50 mM Tris, 0.2% Tween 20 (pH 7.2), and then blocked for 90 min at 25°C with assay buffer (PBS containing 4% BSA and 0.01% Thimerosal, pH 7.2). The plates were washed and 50 μl of assay buffer was added to each along with 50 μl of sample or standard prepared in assay buffer and incubated at 37°C for 2 h. After washing, strepavidin-peroxidase polymer in casein buffer (Research Diagnostics; Mount Pleasant, NJ) was added and incubated at 25°C for 30 min. The plate was washed and 100 μl of substrate (tetramethylbenzidine; DakoCytomation) was added and incubated for 20–30 min. The reaction was stopped with 100 μl of 2 N HCl, and the OD450 (minus OD650) was read on a microplate reader (Molecular Devices). The data was analyzed using a computer program (SoftPro; Molecular Devices). The TNF-, IL-6, IFN-, IL-1, GM-CSF, MIP-2, KC, and IL-8 assays had lower detection limits of 8, 3, 3.9, 1.5, 3, 61, 4, and 4 pg/ml, respectively. Total protein concentrations in supernatants were measured with a commercial reagent (Bio-Rad) with a standard curve constructed using BSA (Sigma-Aldrich).
Histologic analysis
After euthansasia, the anterior chest wall was removed; the trachea was cannulated with an 18-gauge blunt needle; the lungs were inflated in situ with 0.7 ml of 50% (v/v in PBS) OCT medium; the lungs and mediastinum were removed en bloc and embedded in neat OCT, and 8-μm cryosections were acetone-fixed for 3 min at –20°C and stored at –80°C for H&E staining and for immunostaining. A veterinary pathologist (T. O’Neil) evaluated the sections in a blinded fashion for morphologic signs of injury and inflammation. The total area of each lung section with inflammation was measured using a computer-assisted morphometric program (Image-Pro Plus; MediaCybernetics). In the HSF-1-null and control mice exposed to LPS and FRH, the extent of bronchiolar epithelial injury was graded by a blinded observer (J. D. Hasday) using a 0–4 scale where 0 was normal histologic appearance, 1 was loss of cilia, 2 was mild to moderate epithelial cell flattening, 3 was extensive epithelial cell flattening and nuclear degradation, and 4 was denuding of the epithelium with exposure of the underlying basement membrane. For immunohistologic analysis, slides were washed in PBS at 25°C for 20 min and endogenous peroxidase was inactivated by incubating for 10 min with 30% hydrogen peroxide in methanol at 25°C. Nonspecific signal was reduced by sequentially blocking with 5% serum (v/v) in PBS for 30 min, and then with a commercial avidin/biotin blocking kit (Vector Laboratories) according to the manufacturer’s instructions. The blocked slides were sequentially incubated with primary Ab for 1h, a secondary Ab (where indicated) for 30 min, a commercial avidin/biotin peroxidase detection system (Vector Laboratories), and 1 mg/ml diaminobenzidine (Sigma-Aldrich) in 0.02% hydrogen peroxide in PBS. All blocking and subsequent incubations were at 25°C. The blocking serum and primary and secondary Abs used for each cell type were as follows: for macrophages, mouse serum, 1 μg/ml rat anti-mouse Mac-3, and 2.5 μg/ml biotinylated goat anti-rat IgG1/2a HRP (both from BD Pharmingen). For PMN, bovine serum and the 2.5 μg/ml biotinylated rat anti-mouse Gr-1 (BD Pharmingen). For heat shock protein (HSP)72, 5% rabbit serum, 4 μg/ml goat anti-mouse HSP72 (Santa Cruz Biotechnology #SC-1060), and 1/200 dilution of rabbit anti-goat HRP (Sigma-Aldrich #A 5420). Immunostained slides were counterstained with hematoxylin.
Determination of myeloperoxidase activity in tissue
Following euthanasia, the pulmonary circulation was cleared with 10 ml of PBS injected through the left ventricle and drained through an incision in the right atrium. The lungs were dissected free of the trachea and mediastinal contents and homogenized in 1 ml of 50 mM potassium phosphate buffer, pH 6, containing 5 g/L hexadecyltrimethylammonium bromide, passed through a 21-gauge needle three times, sonicated, then subjected to three freeze-thaw cycles. Myeloperoxidase was measured in a 150-μl reaction containing 0.167 mg/ml o-dianisidine (Sigma-Aldrich) and 0.0005% hydrogen peroxide, 50 mM potassium phosphate buffer, pH 6, and 5 μl of sample. Following incubation at room temperature, the OD450 was measured and the myeloperoxidase activity was determined from a standard curve (Sigma-Aldrich). Total protein concentration was measured using Coomassie blue-based assay (Bio-Rad) and a BSA standard curve. Myeloperoxidase activity was standardized to protein concentration.
Epithelial cell culture models
A549 cells were obtained from the American Type Culture Collection and 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 (Invitrogen Life Technologies), pH 7.3 (CRPMI) and containing 10% defined FBS (HyClone) at 37°C in 5% CO2-enriched air. To measure IL-8 secretion, 1 x 105 A549 cells were plated in 1 ml of RPMI 1640/10% FBS in 24-well culture plates, and monolayers were established overnight. The culture medium was changed and, following a 30-min preincubation at the indicated temperature, 2 ng/ml recombinant human TNF- (R&D Systems) or 2 ng/ml IL-1 was added to the culture medium and incubated for 6 h at 37 or 39.5°C or at 42°C for the initial 2 h followed by 4 h at 37°C. IL-8 levels in the 6-h culture supernatants were measured by ELISA.
Immunoblotting
Cell extracts were prepared in RIPA buffer containing 10 mM Tris-HCl (pH 8.0), 150 mM NaCl, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, protease inhibitors (inhibitor cocktail; Boehringer Mannheim), 50 mM NaF, 20 mM -glycerophosphate, 1 mM sodium vanadate, 2 nM calyculin A, and 10 nM okadaic acid. For immunoblotting, 20 μg of total protein was separated by 7.5% SDS-PAGE; electrostatically transferred to polyvinylidene difluoride membrane (Stratagene); blocked with 5% nonfat dry milk in 25 mM Tris (pH 7.4), 0.5 M NaCl/0.05% (v/v) Tween 20 for 30 min at room temperature; and incubated with 1/10,000 dilution of anti-HSP72 Ab (Santa Cruz Biotechnology) in blocking buffer for 1 h. Bands were detected using 1/5000 dilution of goat anti-rabbit IgG HRP conjugate (Bio-Rad) in blocking buffer for 30 min at room temperature, developed with a chemiluminescence detection system (Renaissance; New England Nuclear), and exposed to x-ray film.
Data analysis
Normal data are presented as mean ± SEs and differences among more than two groups were analyzed by applying a post hoc Fisher protected least-squares difference to a one-way ANOVA and differences between two groups was analyzed by Student’s t test. Survival was analyzed with a Gehan-Wilcoxon test of a Kaplan-Meier plot. All experiments were performed at least two times and tested for reproducibility by two-way ANOVA before combining. The epithelial injury scores in the experiment with HSF-1-null mice and the bacterial loads in necropsied mice were not normal and were analyzed by the Mann-Whitney U test.
Results
Effect of FRH on survival in experimental K. pneumoniae pneumonia
We previously found that housing mice with experimental K. pneumoniae peritonitis at 34.5°C rather than 25°C increased core temperature from 36.5–37°C to 39.5–40°C and improved survival (20). To determine whether mice challenged with bacterial agonists via the airway route have a similar relationship between ambient and core temperatures, we telemetrically monitored core temperature in mice that were housed at either 25°C (euthermic) or 34.5°C (FRH) following i.t. challenge with 50 μg of LPS or inoculated with 750 CFU K. pneumoniae Caroli strain (Fig. 1A). In the LPS-challenged mice, exposure to FRH increased core temperature from 37.2 ± 0.2°C to 39.6 ± 0.2°C and in the i.t. K. pneumoniae-inoculated mice from 36.6 ± 0.10°C to 39.0 ± 0.09°C. To determine whether FRH improves survival in experimental pneumonia, groups of 10 mice were briefly anesthetized, i.t. inoculated with 250 CFU K. pneumoniae Caroli strain, and, upon waking, were placed in either 25 or 34.5°C cages and monitored for survival (Fig. 1B). In contrast with our previous observations that FRH improves survival in a K. pneumoniae peritonitis model (20), FRH failed to improve, and tended to worsen survival in mice with experimental K. pneumoniae pneumonia (p = 0.25). We previously demonstrated that the improved survival conferred by FRH in mice with experimental K. pneumoniae peritonitis was associated with a profound reduction in i.p. and systemic bacterial burden (20). FRH exerted similar effects on intrapulmonary bacterial burden in experimental K. pneumoniae pneumonia (Fig. 1C), reducing the median 72-h-postinoculation pulmonary bacterial burden 400-fold (p < 0.02). Little bacterial dissemination occurred in either euthermic (83 ± 75 CFU per milligram of spleen tissue) or FRH (0.9 ± 0.9 CFU per milligram of spleen tissue by 72 h postinoculation. Furthermore, necropsy analysis revealed that the median pulmonary bacterial burden at time of death was 85-fold lower in the warmer animals (4.68 x 104 vs 4.0 x 106 CFU per milligram of lung tissue, six mice per group; p = 0.016), suggesting that mechanisms other than proliferating bacteria contributed to death in the warmer animals.
FIGURE 1. Effect of FRH on survival and pulmonary pathogen burden in experimental K. pneumoniae pneumonia. A, Groups of four mice were implanted with i.p. thermistors and 1 wk later, were challenged with 50 μg of LPS i.t., (squares) or 750 CFU K. pneumoniae Caroli (circles) and placed at 24°C (euthermic; open symbols) or 34°C (FRH; closed symbols) and body temperature was measured telemetrically. B, Groups of 10 mice were inoculated with 250 CFU Klebsiella pneumoniae i.t., then housed at 24°C or 34°C, and survival was monitored. Duplicate experiments were performed. C, Groups of six euthermic and FRH-exposed mice were sequentially euthanized, and bacterial load in lung tissue was quantified. Mean ± SE. *, p < 0.05 vs euthermic mice.
To further analyze the cause of lethality in FRH-treated mice, we developed an antibiotic-treated K. pneumoniae pneumonia model. This model permits us to examine mechanisms of lethality in the absence of continued bacterial growth. Euthermic mice were i.t. inoculated with 250 CFU K. pneumoniae Caroli strain and, 48 h later, treatment with 28 mg/kg ceftriaxone i.p. per day was initiated. By day 5 of antibiotic treatment, proliferating bacteria were eliminated in lung tissue of euthermic animals (Fig. 2A). However, when mice were cotreated with FRH beginning simultaneously with antibiotic treatment, survival was reduced from 100% in the euthermic mice to 50% in the hyperthermic mice (Fig. 2B). Four of five deaths in the FRH group occurred within 2 days of initiating FRH exposure. At necropsy, the lungs in the FRH-treated mice were edematous and hyperemic. However, proliferating bacteria were undetectable in the lungs of three of five dead FRH-exposed mice, indicating that FRH did not antagonize and may have synergized with the bactericidal activity of the antibiotic. Furthermore, these observations suggest that despite the successful elimination of pathogens in experimental pneumonia, antibiotics failed to improve survival in FRH-treated mice.
FIGURE 2. Effect of FRH on survival in mice with antibiotic-treated pneumonia. A, Mice were inoculated with 250 CFU i.t. K. pneumoniae as described in Fig. 1 and, 2 days later, began treatment with 28 mg of ceftriaxone per killogram i.p. daily or sham-treated with i.p. saline. All mice remained euthermic throughout the experiment. Mean ± SE. *, p < 0.05 vs sham-treated mice. B, Groups of mice were inoculated with 250 CFU i.t. K. pneumoniae and, coincident with initiation of ceftriaxone therapy, mice were either switched to a 34°C cage (FRH) or maintained at 24°C (euthermic) and survival was monitored.
Coexposure to FRH and i.t. LPS causes lethal lung injury
To further elucidate the mechanisms of lung injury and death in mice coexposed to intrapulmonary bacterial pathogens and FRH, we developed an i.t. LPS-challenge model. Such a challenge with 50 μg of LPS was nonlethal in euthermic mice, but caused 50% mortality when administered to mice that were coexposed to FRH, with all of the deaths occurring within 3 days of i.t. LPS challenge (Fig. 3A). The increased mortality in the FRH-treated mice was associated with earlier and greater accumulation of protein in bronchoalveolar lavage fluid (BALF) (Fig. 3B), indicating acceleration and augmentation of LPS-induced pulmonary vascular endothelial injury in the warmer mice. By 48 h post-LPS challenge, BALF protein concentrations were 1.5 times greater in the warmer mice (8.5 ± 0.9 vs 5.6 ± 0.8 μg/ml; p < 0.02), indicating greater loss of pulmonary vascular endothelial barrier in the warmer mice.
FIGURE 3. FRH increases pulmonary vascular injury and reduces survival in mice challenged with i.t. LPS. Mice were challenged with 50 μg of LPS i.t. and housed at 24°C (euthermic) or 34°C (hyperthermic) and survival was measured (A). Groups of six mice were euthanized at each time point and pulmonary vascular leak was analyzed by measuring total BALF protein concentration (B). Mean ± SE. *, p < 0.05 vs euthermic mice.
Interstitial and alveolar inflammation was first evident by 12 h after LPS instillation in the euthermic mice. Subsequent progression of inflammation was accompanied by type II pneumocyte hypertrophy and mucous accumulation. FRH-exposed mice displayed qualitatively similar, but more extensive changes in lung histology compared with euthermic mice at each time point (Fig. 4, A–C). Interestingly, by 48 h after LPS challenge, widespread bronchiolar epithelial necrosis occurred in the FRH-exposed mice, but not in the euthermic animals (Fig. 5).
FIGURE 4. FRH increases the extent of lung consolidation in i.t. LPS-challenged mice. Mice were challenged with 50 μg of LPS i.t. then placed at 24°C or 34°C. Two mice per group were euthanized at each time point, the lungs were inflated with 0.7 ml of 50% (v/v) OCT, and 5-μm cryosections were stained with H&E and analyzed for area of consolidation using morphometric analysis software. Representative 20x photomicrographs of euthermic (A) and hyperthermic (B) mice 48 h after LPS challenge and means of all data (C) are shown.
FIGURE 5. FRH and i.t. LPS challenge causes bronchiolar epithelial necrosis. Mice were challenged with 50 μg of LPS i.t. then placed at 24°C (euthermic; B–C) or 34°C (hyperthermic; D–F). Two mice per group were euthanized per time point, the lungs were inflated with 0.7 ml of 50% (v/v) OCT, and 5-μm cryosections were stained with H&E. Representative photomicrographs (x200) of an untreated control (A), euthermic mice 36 h (B) and 48 h (C) after LPS challenge, and hyperthermic mice 24 h (D), 36 h (E), and 48 h (F) after LPS challenge. G, H, and I show high power views of A, C, and F, respectively. Note the loss of cilia and loss of distinct nuclei (arrows) in I.
Coexposure to FRH augments PMN accumulation within the LPS-challenged mouse lung
Exposing LPS-challenged mice to FRH augmented the accumulation of PMN in BALF 3-fold 48 h after i.t. LPS challenge from 1.66 ± 0.49 x 106 cells in the euthermic group to 4.1 ± 0.55 x 106 cells in the FRH group (Fig. 6A; p = 0.005). Microscopic analysis of lung tissue immunostained for the PMN marker GR-1 demonstrated PMN accumulation within inflammatory exudates of both groups of LPS-challenged mice; however, these exudates were more extensive in the FRH-exposed than the euthermic mice (Fig. 6, B and C). By 24 h after LPS challenge, myeloperoxidase activity in lung tissue was 2-fold higher in the FRH than euthermic mice (2 ± 0.23 vs 1.07 ± 0.24 myeloperoxidase units per microgram of protein; p < 0.02), providing further evidence of augmented pulmonary PMN accumulation in the warmer mice.
FIGURE 6. FRH augments pulmonary accumulation of PMN after i.t. LPS. Mice were challenged with 50 μg of LPS i.t. then placed at 24°C (euthermic) or 34°C (hyperthermic). Mice were euthanized at the indicated times; BALF was collected and analyzed for total PMN number (A). Mean ± SE; n = 6. *, p < 0.05 vs euthermic. Cryosections from hyperthemic (B) and euthermic (C) mice 48 h after i.t. LPS challenge were immunostained for the PMN marker, GR-1.
We previously demonstrated that FRH increased both pulmonary PMN burden and expression of the ELR+ CXC chemokines, KC, LIX/CXCL5, and MIP-2, and G-CSF in mice exposed to lethal hyperoxia (24). To determine whether FRH might also augment expression of these cytokines during Gram-negative bacterial pneumonia, we sequentially analyzed the cytokine composition of BALF from euthermic and FRH-exposed mice after i.t. LPS challenge (Fig. 7). BALF GM-CSF concentrations were 10-fold greater (253.3 ± 19.2 pg/ml vs 25.5 ± 22.2 pg/ml) in the FRH-exposed mice compared with euthermic animals 6 h after LPS challenge and IL-1 levels were increased 4.8-fold (382 ± 44.2 vs 78.9 ± 47.3 pg/ml) 24 h after LPS challenge. The kinetics of MIP-2 and KC expression were similar. Accumulation of both chemokines in BALF was evident 6 h after LPS challenge and reached comparable levels in the euthermic and hyperthermic mice, but a second peak of expression for each of these cytokines at 24 h post-LPS challenge only occurred in the hyperthermic mice. KC levels were 4.8-fold higher (1366 ± 239 vs 284 ± 106 pg/ml) and MIP-2 levels 2.3-fold higher (417 ± 64 vs 181 ± 67 pg/ml) in the FRH-exposed mice 24 h after LPS challenge compared with euthermic controls at the same time point. The BALF concentrations of TNF- and IFN- were similarly elevated in both groups (data not shown).
FIGURE 7. FRH augments BALF cytokine levels in i.t. LPS-challenged mice. Mice were challenged with 50 μg of LPS i.t., exposed to 24°C (euthermic) or 34°C (hyperthermic) ambient temperature. Groups of six mice were euthanized at each time point and lung lavage was analyzed for cytokine expression by ELISA. Mean ± SE. *, p < 0.05 vs euthermic animals.
Heat shock augments IL-8 secretion by TNF-- and IL-1-stimulated human lung epithelial cells
To determine whether FRH directly augments generation of ELR+ CXC chemokines by lung epithelium we analyzed the effect of in vitro FRH exposure on IL-8 expression by the lung alveolar epithelial-like cell line, A549 (Fig. 8, A and B). Because LPS did not directly activate IL-8 release in A549 cells (data not shown), we stimulated A549 cells with TNF- (Fig. 8A) or IL-1 (Fig. 8B), each of which is an early LPS-induced mediator which was elevated in lung lavage after i.t. LPS challenge. A549 cells released comparable levels of IL-8 at 37 and 39.5°C following stimulation with TNF- (1057 ± 35.7 pg per 106 cells at 37°C vs 862 ± 42.9 pg per 106 cells at 39.5°C) and IL-1 (2700 ± 209 vs 2414 ± 90 pg per 106 cells). In contrast, exposing A549 cells to 42°C for 2 h followed by 4 h incubation at 37°C augmented IL-8 expression 2-fold (2061 ± 120 pg per 106 cells) in TNF--stimulated cells and 2.3-fold (6513 ± 693 pg per 106 cells) in IL-1-stimulated cells compared with cells cultured at 37°C. Although exposing A549 cells to 42°C for 2 h stimulated expression of the inducible HSP, HSP72, exposure to 39.5°C for 6 h failed to activate a measurable increase in HSP72 protein levels (Fig. 8C), suggesting that heat shock may contribute to IL-8 expression.
FIGURE 8. Heat-shock augments TNF-- and IL-1-induced IL-8 expression in A549 cell culture. A549 cells (2 x 105 cells per milliliter of RPMI 1640/10% FBS) were incubated for 6 h with 10 ng/ml recombinant human TNF- (A) or 10 ng/ml IL-1 (B) at 37 or 39.5°C, or were exposed to 42°C for 2 h and switched to 37°C for 4 h (HS), and IL-8 secretion was measured by ELISA. *, p < 0.05 vs 37°C and vs 39.5°C. C, A549 cells were incubated for 4 h at 37°C (lane 1) or 39.5°C (lane 2) or were heated to 42°C for 2 h, then incubated for an additional 2 h at 37°C (HS; lane 3) and lysates were immunoblotted for HSP72 (arrow). A nonspecific band appeared at 50 kDa (arrowhead).
Although direct exposure of A549 cells to 39.5°C in vitro failed to activate a detectable heat shock response, Ostberg et al. (27) previously reported that exposing intact mice to a core temperature of 39.5°C for 6 h activates heat shock response in several organs, including the lungs. To determine whether FRH activates heat shock in the i.t. LPS-challenged mouse lung and identify the cellular expression pattern, we analyzed lung cryosections that were immunostained for anti-HSP72 (Fig. 9). Coexposure to FRH and LPS for 24 h induced expression of HSP72 primarily within the bronchiolar and bronchial epithelium (Fig. 9A, see arrows). In contrast, only minimal HSP72 expression was evident in the lungs of LPS-challenged, euthermic mice.
FIGURE 9. FRH induces expression of HSP72 in bronchiolar epithelium. Mice were challenged with 50 μg of LPS i.t. then placed at 24°C (euthermic) or 34°C (hyperthermic) for 24 h. Cryosections from hyperthermic (A) and euthermic (B) mice 24 h after i.t. LPS challenge or untreated control mice (C) were immunostained for HSP72. Arrows in A indicate areas of increased bronchiolar epithelial HSP72 staining.
To determine whether the heat-shock response activated in mice exposed to FRH contributes to the pattern of injury induced by coexposure to FRH and i.t. LPS, we compared the histological changes within the lung of HSF-1-null mice and their heterozygous littermates 24 h after exposure to 50 μg of LPS i.t. and FRH (Fig. 10). The heterozygous mice showed an injury pattern similar to that seen in the CD-1 mice exposed to LPS and FRH, with diffuse bronchiolar epithelial cell necrosis. In contrast, the bronchiolar epithelium remained well-preserved in the HSF-1-null mice despite exposure to the same combination of LPS and FRH.
FIGURE 10. HSF-1-null mice are resistant to bronchiolar epithelial necrosis after exposure to LPS and FRH. HSF-1-null mice (A) and their heterozygous littermates (B) were challenged with 50 μg of LPS i.t. then placed at 34°C. Four mice per group were euthanized 24 h later, the lungs were inflated with 0.8 ml of 50% (v/v) OCT, and 5-μm cryosections were stained with H&E. Representative x100 photomicrographs of HSF-1-null (A) and heterozygous (B) mice are shown. C, The summary data of arbitrary bronchiolar epithelial injury scores of the two groups (mean ± SE; p = 0.011).
Discussion
We have previously shown that increasing core temperature from basal to febrile levels improves survival in K. pneumoniae peritonitis by accelerating pathogen clearance (20). However, death still occurred in some of the warmer mice despite reduced pathogen loads, suggesting that the host response rather than the pathogen itself may contribute to death in the warmer mice. Nonetheless, the improved survival in the FRH-treated mice demonstrates that in the peritonitis model, the beneficial effects of FRH on pathogen clearance outweigh its potential harmful effects. In the present study, we extended these observations by analyzing the effect of FRH on immune responses to a bilateral multilobar Gram-negative bacterial pneumonia, an antibiotic-treated pneumonia, and i.t. LPS. We found that FRH failed to improve survival in the K. pneumoniae pneumonia model despite affecting a profound reduction in pulmonary bacterial burden. We showed that antibiotic-treated pneumonia and i.t. LPS challenge, each of which was nonlethal in euthermic mice, caused 50 and 45% mortality, respectively, when accompanied by FRH. In the intratrachal LPS model, exposure to FRH increased pulmonary PMN accumulation, pulmonary vascular endothelial leak, and bronchiolar epithelial injury. Hyperthermia also augmented IL-1, MIP-2, KC, and GM-CSF secretion into the bronchoalveolar compartment. These data suggest that, in the setting of diffuse bacterial pneumonia, FRH increases generation of proinflammatory cytokines, increases recruitment of PMN, augments immune-mediated pulmonary endothelial and epithelial injury, and increases mortality.
Our temperature-controlled mouse model was designed to isolate the effects of febrile temperature itself from the processes that generate fever. The effects of FRH on pulmonary inflammation and epithelial injury were not due to direct airway warming, as the temperature of the inhaled air (34°C) was below normal core temperature (37°C). The use of anesthesia during the inoculation procedure avoided the effects of physical stress and provided more precise control of the amount of inoculum entering the lungs. Although anesthetic agents may influence the acute-phase response, three lines of evidence suggest that the anesthetics used in this study did not interfere with the studied immune parameters. First, previous studies in our laboratory found that use of tribromoethanol had no effect on LPS-induced cytokine expression in euthermic mice, nor did anesthesia modify the effects of hyperthermia on LPS-induced cytokine expression (28). Second, in the present study, the euthermic and hyperthemic mice were treated with the same anesthesia protocol. Third, no toxic effects of anesthesiaadministration were found in temperature-controlled mice mock challenged with sterile saline i.t.
We previously reported that exposing mice to FRH activates expression of circulating G-CSF, expansion of the circulating PMN pool, and, in mice exposed to hyperoxia for 24 h, enhances bronchoalveolar expression of the ELR+ CXC chemokines, KC, LIX, and MIP-2 (24). In the present study, i.t. LPS induced early expression of KC and MIP-2 in BALF evident at 6 h following LPS challenge, which was not modified by coexposure to FRH. However, a second peak of BALF KC and MIP-2 occurred 24 h after LPS challenge, but only in the mice exposed to FRH. Exposure to FRH stimulated a similar increase in BALF IL-1 levels 24 h post-LPS challenge that temporally correlated with the augmented chemokine expression in the warmer mice. In contrast with the effects of FRH on chemokine expression, coexposure to FRH augmented early GM-CSF expression evident at 6 h post-LPS challenge by 10-fold, but did not stimulate subsequent GM-CSF expression. Furthermore, while the effects of FRH on KC and MIP-2 expression in LPS-challenged mice in the present study were similar to those previously found in hyperoxia-exposed mice (24), FRH failed to augment BALF GM-CSF or IL-1 levels in the hyperoxia mouse model (24). Although the proximal signaling events that mediate pulmonary oxygen toxicity are not well understood, the susceptibility of TLR4-defective, LPS-resistant C3H/HeJ mice to hyperoxia-induced lung inflammation and injury (29) suggests that the lung injury induced by LPS and by hyperoxia are triggered by distinct signaling events. Collectively, these data suggest that FRH exerts specific effects on a subset of cytokine genes and that the pattern of gene expression is determined in part by the proximal injurious agent. Interestingly, the pattern of augmented cytokine generation in these models suggests a coordinated pattern of biologic actions through which fever may augment PMN activation and delivery to sites of infection and injury. We previously found that the augmented PMN accumulation in the lungs of mice exposed to hyperoxia was completely blocked by immunoblockade of the common receptor for CXC chemokines (24). However, in the current study, PMN content continued to increase between 24 and 48 h after LPS challenge in the FRH-exposed mice despite a peak and subsequent decline in BALF levels of MIP-2 and KC over this same time. These data suggest that additional chemotactic factors (e.g., complement protein C5a fragment or leukotriene B4) may contribute to the late PMN accumulation in the warmer mice. Furthermore, the observation by Bernheim et al. (18) that exposure to FRH causes a similar accumulation of granulocytes at sites of Gram-negative infection in a more primitive animal, the desert iguana Diposuaurus dorsalis, suggests that these processes may be phylogenetically old and evolutionarily conserved.
We showed that exposure to core temperature of 39.5°C was sufficient to stimulate an increase in HSP72 immunostaining in the LPS-challenged mouse lung. The anti-HSP72 Ab, which specifically recognizes the heat-shock-inducible HSP70 family member, failed to immunostain lung tissue from euthermic mice. These results are consistent with those of Ostberg et al. (27), in which they showed that exposing mice to 39.5–40°C for 6 h induced expression of HSP72 and HSP110 in several organs, including lung. Furthermore, the thermal threshold for heat shock may be reduced further by inflammatory mediators that may be present in the infected or injured lung, including arachidonic acid (30), type I IFNs (31), and oxidants (32). Collectively, these data support the concept that heat shock activation may occur during febrile illnesses.
We have previously shown that FRH (39.5°C) causes only modest augmentation in TNF--induced expression of the human ELR+ CXC chemokine, IL-8, and GM-CSF in cultured human pulmonary artery endothelial cells (33). In the present study, in vitro exposure to FRH (39.5°C) failed to augment detectable IL-8 or GM-CSF expression in A549 cultured epithelial cells stimulated by either TNF- or IL-1. However, heating the same cells to 42°C for 2 h augmented TNF-- and IL-1-induced expression of IL-8 and GM-CSF and induced a heat shock response as indicated by the appearance of HSP72 protein. TNF- and IL-1 use distinct proximal signaling pathways that converge on NF-B, a transcription factor critical for IL-8 and GM-CSF gene activation. However, several studies have demonstrated that heat shock blocks NF-B activation (34). Yoo et al. (35) reported that pre-exposure to heat shock reduced NF-B activation and IL-8 secretion in BEAS-2B and A549 cultured respiratory epithelial cells by inhibiting IKK and stabilizing IB. In contrast with the coincident exposure of A549 cells to heat shock and proinflammatory stimuli analyzed in the present study, the studies showing an inhibitory effect of heat shock on NF-B activation used a protocol in which the inflammatory stimulus was added after heat shock had been induced. On the other hand, Bowman et al. (36) showed that exposing cultured human keratinocytes to 54°C for 4 s activated expression of both HSP72 and IL-8. These studies are limited to showing an association between heat shock activation and expression of ELR+ CXC chemokines. Additional studies establishing a regulatory role for heat shock pathways in activation of the ELR+ CXC chemokine gene family are ongoing in our laboratory.
It is notable that in the present study, exposing mice to FRH augmented BALF IL-1 levels 24 h after LPS challenge. In contrast, we and others have shown that direct exposure of macrophages in vitro to FRH (39.5°C) (37) or heat shock (42°C) temperature (38) attenuates IL-1 expression, in part by inhibiting C/EBP-mediated transcription (38). Possible explanations for the discrepancy between the in vitro and in vivo effects of heating include: 1) heat shock exerts different effects on IL-1 transcription in lung macrophages in vivo and monocytes and macrophages in vitro; 2) the augmentation of IL-1 secretion in vivo is mediated through posttranscriptional effects of heat shock; and 3) that cells other than macrophages secrete the excess IL-1 in the warmer mice. The temporal correlation between the rise in BALF IL-1 levels and PMN influx and the known capacity of PMN to express IL-1 (39) implicates this cell as one potential source of excess IL-1 in the FRH-exposed mice. However, definitive identification of the cellular source of excess IL-1 in the FRH-exposed, LPS-challenged lung awaits the result of ongoing studies in our laboratory.
Although several studies by other groups have shown that hyperthermia protects against pulmonary injury (40, 41, 42, 43), the heating protocols used in these studies achieved higher temperatures, 40.5–42°C, for shorter periods, 15–20 min, and allowed for recovery of up to 18 h before exposure to injurious agents. In contrast with the heat shock preconditioning protocols used in these studies, the warming protocol used in the present study was designed to model fever associated with naturally occurring infections. Using a similar coexposure protocol, Suzuki et al. (44) recently reported that warming rabbits 2°C above their basal core temperature augmented lung injury induced by high tidal volume mechanical ventilation in vivo and in isolated lung preparations perfused with 2% blood ex vivo. However, because lung injury in this model was evident after only 2 h of hyperthermia and mechanical ventilation and occurred in the excised lungs despite the absence of PMN in the perfusate, additional mechanisms may have contributed to lung injury in the warmer mice in this study. We previously demonstrated that FRH (39.5°C) augments TNF--induced paracellular leak in cultured human pulmonary artery endothelial cells in vitro (33), thereby demonstrating another potential effect of FRH that may contribute to lung injury.
Whereas, cytokine expression, PMN recruitment, and pulmonary vascular endothelial leak were stimulated by LPS challenge in both euthermic and hyperthermic mice, albeit to higher levels in the warmer mice, extensive bronchiolar epithelial necrosis occurred only in the hyperthermic mice. Expression of HSP72, an indicator of the heat shock response, was also evident in the bronchial and bronchiolar epithelium of the warmer mice. Although the temporal and spatial correlation between epithelial necrosis and HSP72 expression suggests a possible mechanistic link, we have not yet generated evidence that heat shock itself leads to bronchiolar epithelial necrosis in the FRH-exposed, LPS-challenged lung. Although most studies have concluded that induction of thermotolerance or overexpression of HSP72 blocks apoptosis (45, 46, 47), elicitation of the heat shock state preceded exposure to the apoptotic stimulus in these studies. In contrast, in febrile illnesses, HSPs are expressed during the inflammatory response. Furthermore, Buzzard et al. (47) reported that overexpression of HSP72 in mouse embryonic fibroblasts exerted anti-apoptotic action against some (e.g., TNF- and heat) but not all (e.g., ionizing radiation) apoptotic stimuli. Similarly, Gabai et al. (46) reported that HSP72 overexpression in human fibroblasts blocked early, Bid-dependent apoptosis, but not subsequent Bid-independent apoptosis. It is also possible that the bronchiolar epithelial necrosis that occurred in the warmer mice is not caused directly by the expression of HSP genes within the epithelium. These animals were not only subjected to a greater intrapulmonary PMN burden, but they also were exposed to higher levels of the PMN activating cytokines, GM-CSF, KC, MIP-2, and IL-1. Furthermore, FRH itself directly enhances PMN generation of reactive oxygen intermediates and NO responsiveness to LPS (48). In fact, we previously reported that immunoblockade of the common receptor for KC and MIP-2 in hyperoxia-exposed mice partially reversed the accelerated death caused by coexposure to FRH (24).
In summary, we have shown that the same exposure to FRH that was protective in K. pneumoniae peritonitis (20), failed to improve survival in pneumonia with the same pathogen. These results suggest that FRH exerts profound immunomodulating effects during infections, and that the consequences of fever are determined by a balance between accelerated pathogen clearance and augmented collateral tissue injury. The disparate effect of FRH on survival in bacterial peritonitis and pneumonia indicates that the ultimate effects of fever are determined by the site of infection or injury as well, and that the lung may be particularly susceptible to augmented tissue injury during fever. These results show that there is a need for future studies of fever management in well-defined patient populations.
Disclosures
The authors have no financial conflict of interest.
Acknowledgments
We thank Dr. Matthew Fenton for his thoughtful review of the manuscript.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 This work was supported by National Institutes of Health Grants HL69057, GM066855, and AI42117 and a Veterans Affairs Merit Review Award.
2 Address correspondence and reprint requests to Dr. Jeffrey D. Hasday, Baltimore Veterans Administration Medical Center, Room 3D122, 10 North Greene Street, Baltimore, MD 21201. E-mail address: jhasday{at}umaryland.edu
3 Abbreviations used in this paper: FRH, febrile-range hyperthermia; HSP, heat shock protein; i.t., intratracheal; PMN, polymorphonuclear cell; BALF, bronchoalveolar lavage fluid.
Received for publication August 20, 2004. Accepted for publication December 27, 2004.
References
Kluger, M. J., W. Kozac, C. A. Conn, L. R. Leon, D. Soszynski. 1996. The adaptive value of fever. P. A. Mackowiak, ed. Fever: Basic Mechanisms and Management 255. Raven Press, New York.
Doran, T. F., C. DeAngelis, R. A. Baumgardner, E. D. Mellits. 1989. Acetaminophen: more harm than good for chicken pox?. J. Pediatr. 114:1045.
Graham, N. M. H., C. J. Burrell, R. M. Douglas, P. Debelle, L. Davies. 1990. Adverse effects of aspirin, acetaminophen, and ibuprofen on immune function, viral shedding, and clinical status in rhinovirus-infected volunteers. J. Infect. Dis. 162:1277
Stanley, E. D., G. G. Jackson, C. Panusarn, M. Rubenis, V. Dirda. 1975. Increased viral shedding with aspirin treatment of rhinovirus infection. J. Am. Med. Assoc. 231:1248.
Mackowiak, P. A., S. S. Wasserman, M. M. Levine. 1992. An analysis of the quantitative relationship between oral temperature and severity of illness in experimental shigellosis. J. Infect. Dis. 166:1181.
Bryant, R. E., A. F. Hood, C. E. Hood, M. G. Koenig. 1971. Factors affecting mortality of gram-negative rod bacteremia. Arch. Intern. Med. 127:120.
Mackowiak, P. A., R. H. Browne, P. M. Southern, Jr, J. W. Smith. 1980. Polymicrobial sepsis: an analysis of 184 cases using log linear models. Am. J. Med. Sci. 280:73.
Weinstein, M. P., P. B. Iannini, C. W. Stratton, T. C. Eickhoff. 1978. Spontaneous bacterial peritonitis: a review of 28 cases with emphasis on improved survival and factors influencing prognosis. Am. J. Med. 64:592.
Hoefs, J. C., H. N. Canawati, F. L. Sapico, R. R. Hopkins, J. Weiner, J. Z. Montgomerie. 1982. Spontaneous bacterial peritonitis. Hepatology 2:399.
Ahkee, S., L. Srinath, J. Ramirez. 1997. Community-acquired pneumonia in the elderly: association of mortality with lack of fever and leukocytosis. South Med. J. 90:296
Hasday, J. D., K. D. Fairchild, C. Shanholtz. 2000. The role of fever in the infected hosts. Microbes Infect. 2:1891
Vaughn, L. K., H. A. Bernheim, M. J. Kluger. 1974. Fever in the lizard Dipsosaurus dorsalis. Nature 252:473.
Armstrong, C.. 1942. Some recent research in the field of neurotropic viruses with especial reference to lymphocytic choriomeningitis and herpes simplex. Milit. Surg. 91:129.
Bell, J. F., G. F. Moore. 1974. Effects of high ambient temperature on various stages of rabies virus infection in mice. Infect. Immun. 10:510.
Covert, J. B., W. W. Reynolds. 1977. Survival value of fever in fish. Nature 267:43
Bernheim, H. A., P. T. Bodel, P. W. Askenase, E. Atkins. 1978. Effects of fever on host defence mechanisms after infection in the lizard Dipsosaurus dorsalis. Br. J. Exp. Pathol. 59:76
Bernstein, I. L., H. S. Boushey, R. M. Cherniack, J. N. Fink, J. D. Fulmer, E. J. Goetzl, R. A. Goldstien, S. S. Hurd, J. S. Ram, L. M. Lichtenstein, et al 1985. Summary and recommendations of a workshop in the investigative use of fiberoptic bronchoscopy and bronchoalveolar lavage in asthmatics. Am. Rev. Respir. Dis. 132:180.
Jiang, Q., A. S. Cross, I. S. Singh, T. T. Chem, R. M. Viscardi, J. D. Hasday. 2000. Febrile core temperature is essential for optimal host defense in bacterial peritonitis. Infect. Immun. 68:1265.[Abstract/Free Full Text]
Kluger, M. J., D. H. Ringler, M. R. Anver. 1975. Fever and survival. Science 188:166.
Schmidt, J. R., A. F. Rasmussen, Jr. 1960. The influence of environmental temperature on the course of experimental herpes simplex infection. J. Infect. Dis. 107:356
Ellingson, H. V., P. F. Clark. 1942. The influence of artificial fever on mechanisms of resistance. J. Immunol. 43:65
.
Hasday, J., A. Garrison, I. Singh, T. Standiford, G. Ellis, S. Rao, J.-R. He, P. Rice, M. Frank, S. Goldblum, R. Viscardi. 2003. Febrile-range hyperthermia augments pulmonary neutrophil recruitment and amplifies pulmonary oxygen toxicity. Am. J. Pathol. 162:2005
Xiao, X., X. Zuo, A. A. Davis, D. R. McMillan, B. B. Curry, J. A. Richardson, I. J. Benjamin. 1999. HSF1 is required for extra-embryonic development, postnatal growth, and protection during inflammatory response in mice. EMBO J. 18:5943.
Keane-Myers, A. M., W. C. Gause, F. D. Finkelman, X. D. Xhou, M. Wills-Karp. 1998. Development of murine allergic asthma is dependent upon B7-2 costimulation. J. Immunol. 160:1036.
Ostberg, J. R., K. C. Kaplan, E. A. Repasky. 2002. Induction of stress proteins in a panel of mouse tissues by fever-range whole body hyperthermia. Int. J. Hyperthermia 18:552.
Jiang, Q., L. DeTolla, I. Kalvakolanu, B. Fitzgerald, J. D. Hasday. 1999. Fever upregulates expression of pyrogenic cytokines in endotoxin-challenged mice. Am. J. Physiol. 276:R1653.
Johnston, C. J., B. R. Stripp, B. Piedbeouf, T. W. Wright, G. W. Mango, C. K. Reed, J. N. Finkelstein. 1998. Inflammatory and epithelial responses in mouse strains that differ in sensitivity to hyperoxic injury. Exp. Lung Res. 24:189.
Jurivich, D. A., L. Sistonen, K. D. Sarge, R. I. Morimoto. 1994. Arachidonate is a potent modulator of human heat shock gene transcription. Proc. Natl. Acad. Sci. USA 91:2280.
Bowman, P. D., S. T. Schuschereba, D. F. Lawlor, G. R. Gilligan, J. R. Mata, D. R. DeBaere. 1997. Survival of human epidermal keratinocytes after short-duration high temperature: synthesis of HSP70 and IL-8. Am. J. Physiol. 272:C1988.
Fairchild, K. D., R. M. Viscardi, L. Hester, I. S. Singh, J. D. Hasday. 2000. Effects of hypothermia and hyperthermia on cytokine production by cultured human mononuclear phagocytes from adults and newborns. J. Interferon Cytokine Res. 20:1049
Gabai, V. L., K. Mabuchi, D. D. Mosser, M. Y. Sherman. 2002. Hsp72 and stress kinase c-jun N-terminal kinase regulate the bid-dependent pathway in tumor necrosis factor-induced apoptosis. Mol. Cell Biol. 22:3415.
Buzzard, K. A., A. J. Giaccia, M. Killender, R. L. Anderson. 1998. Heat shock protein 72 modulates pathways of stress-induced apoptosis. J. Biol. Chem. 273:17147
Rosenspire, A. J., A. L. Kindzelskii, H. R. Petty. 2002. Cutting edge:fever-associated temperatures enhance neutrophil responses to lipopolysaccharide: a potential mechanism involving cell metabolism. J. Immunol. 169:5396(Penelope Rice, Erica Mart)
We previously demonstrated that exposure to febrile-range hyperthermia (FRH) accelerates pathogen clearance and increases survival in murine experimental Klebsiella pneumoniae peritonitis. However, FRH accelerates lethal lung injury in a mouse model of pulmonary oxygen toxicity, suggesting that the lung may be particularly susceptible to injurious effects of FRH. In the present study, we tested the hypothesis that, in contrast with the salutary effect of FRH in Gram-negative peritonitis, FRH would be detrimental in multilobar Gram-negative pneumonia. Using a conscious, temperature-clamped mouse model and intratracheal inoculation with K. pneumoniae Caroli strain, we showed that FRH tended to reduce survival despite reducing the 3 day-postinoculation pulmonary pathogen burden by 400-fold. We showed that antibiotic treatment rescued the euthermic mice, but did not reduce lethality in the FRH mice. Using an intratracheal bacterial endotoxin LPS challenge model, we found that the reduced survival in FRH-treated mice was accompanied by increased pulmonary vascular endothelial injury, enhanced pulmonary accumulation of neutrophils, increased levels of IL-1, MIP-2/CXCL213, GM-CSF, and KC/CXCL1 in the bronchoalveolar lavage fluid, and bronchiolar epithelial necrosis. These results suggest that FRH enhances innate host defense against infection, in part, by augmenting polymorphonuclear cell delivery to the site of infection. The ultimate effect of FRH is determined by the balance between accelerated pathogen clearance and collateral tissue injury, which is determined, in part, by the site of infection.
Introduction
Fever is generally regarded as a protective response to infection (1). Retrospective clinical studies demonstrate an association between the presence of fever and reduced length of viral illnesses (2, 3, 4, 5) or improved survival in bacterial infections (6, 7, 8, 9, 10). However, this association is lost in patients with higher disease acuity (11). In two retrospective clinical studies of patients with bacteremia and sepsis, treatment with the antipyretic agent, acetaminophen, was associated with improved survival, although actual core temperatures were not mentioned in the report (12, 13). Studies in animal models in which febrile-range hyperthermia (FRH)3 was achieved by increasing ambient temperature generally support the concept that a temperature increase to febrile levels is protective during infections (14, 15, 16, 17, 18, 19, 20, 21, 22). When the desert lizard, Diposaurus dorsalis, is inoculated s.c. with the Gram-negative pathogen, Aeromonas hydrophila, it increases its core temperature by seeking a warmer environment (14, 18, 21). In this model, a 2°C increase in core temperature improves survival, accelerates pathogen clearance, and augments accumulation of granulocytes at the inoculation site (18). Similarly, we have shown that housing mice with experimental Klebsiella pneumoniae peritonitis at 34.5°C rather than 23°C increased core temperature from 37 to 39.7°C, reduced i.p. bacterial load by five orders of magnitude, and increased survival from 0 to 50% (20). In that study, in vitro K. pneumoniae proliferation rates were comparable in 37 and 39.5°C bacterial cultures, indicating that the accelerated bacterial clearance in the warmer animals was mediated through enhanced host defense rather than direct bactericidal effects of FRH. Interestingly, while death in the euthermic mice was associated with overwhelming infection, death in the hyperthermic mice occurred in the face of much lower bacterial burdens. The discordance between lethality and bacterial burden in the warmer mice suggested a greater contribution from the host immune response to lethality in the presence of FRH. A similar discordance between bacterial load and survival was found by Ellingson et al. (23) in rabbits with experimental Streptococcus pneumoniae. Based on these studies, we reasoned that the ultimate effect of fever would be determined by a dynamic balance between accelerated pathogen clearance and augmented collateral host tissue injury, and would depend, in part, on the nature of the organ system involved. Because of its delicate anatomic structure and its critical gas exchange function, the lower respiratory tract might be less tolerant of the immunostimulatory effects of FRH. We recently found that lethal lung inflammation and injury is augmented by FRH in a mouse model of pulmonary oxygen toxicity (24), illustrating the potential harm of FRH-augmented lung injury. However, hyperoxia is an artificial injurious agent that stimulates injury in the absence of a replicating pathogen.
In the present study, we tested the hypothesis that, in contrast with its salutary effect in Gram-negative peritonitis (20), FRH would be detrimental in diffuse Gram-negative pneumonia because the augmented collateral lung injury would offset any enhancement of pathogen elimination. We used a conscious, temperature-clamped mouse model to elucidate the effects of FRH on the outcome of K. pneumoniae pneumonia, antibiotic-treated K. pneumoniae pneumonia, and intratracheal (i.t.) bacterial endotoxin LPS challenge. To further analyze the effects of FRH on collateral tissue injury, we focused on the potential contribution of polymorphonuclear cells (PMNs) and the role of CXC chemokines and GM-CSF in the FRH-augmented lung injury.
Materials and Methods
Mice and reagents
Eight- to 10-wk-old male outbred CD-1 mice, weighing 25–30 g were purchased from Harlan Sprague Dawley, housed in the Baltimore Veterans Administration Medical Center animal facility under American Association of Laboratory Animal Care-approved conditions and under the supervision of a full-time veterinarian. All animals were used within 4 wk of delivery. Heat shock factor-1 (HSF-1)-null mice have been previously characterized (25) and a colony established at the University of Maryland. Homozygous HSF-1-null males were bred with heterozygous females and the genotypes determined from PCR analysis of tail snip DNA as previously described (25). LPS prepared by trichloroacetic acid extraction from Escherichia coli O111:B4 and 2,2,2-tribromoethyl were purchased from Sigma-Aldrich. All protocols were approved by the Insitutional Animal Care and Use Committee of the University of Maryland, Baltimore.
Bacterial culture
K. pneumoniae 1:K2 strain B5055 and the Caroli substrain of B5055 were obtained from Dr. I. Orskov and Dr. F. Orskov (State Serum Institute, Copenhagen, Denmark) and stored as frozen glycerol stocks. The bacteria were cultured in tryptic soy broth (Sigma-Aldrich) at 37°C and growth was monitored by measuring OD650. Bacteria were harvested during logarithmic growth, washed three times with sterile PBS (pH 7.2), adjusted to 0.3 OD units (1 x 108 bacteria per milliliter), and used to inoculate mice. Inoculum size was confirmed by enumerating colony growth on MacKonkey agar.
Temperature clamping, experimental pneumonia, and i.t. administration of LPS
Mice were adapted to standard plastic cages for at least 4 days before study. To avoid the influence of diurnal cycling, all experiments were started at approximately the same time each day (between 8:00 and 10:00 a.m.). To measure core temperature, 1 wk before a planned experiment, one sentinel mouse from each experimental group was anesthetized with 250 mg/kg tribromoethanol (Avertin) i.p. A sterilized Mini Mitter 1.05g in vivo temperature sensor, model 100-0035, was freely placed into the peritoneal cavity, the peritoneum and abdominal musculature were closed with a single layer of sutures, and the skin closed using surgical glue (Vet-Bond). Euthermia was maintained by holding the mice in a 36.5°C infant isolette until they awakened. The mice received one dose of 2.5 mg/kg buprenorphine analgesia s.c. immediately after surgery and food and water were provided ad libitum. The signal from the temperature sensors was continuously monitored using the Mini Mitter Automated Data Acquisition System.
On the day of inoculation, mice were anesthetized with 250 mg/kg Avertin i.p. and placed supine at a 45° angle. The tongues were gently extended using forceps, 50 μl of K. pneumoniae inoculum or 1 mg/ml LPS in sterile PBS was pipetted gently into the posterior pharynx, and the mice held in this position until aspiration was confirmed as described by Keane-Myers et al. (26). In preliminary experiments, we showed that Evans blue dye administered using this method stained the surfaces of all lobes of both lungs with only small amounts swallowed or retained in the trachea. To prevent anesthetic-induced hypothermia, euthermia was maintained in a 36.5°C isolette until the mice awakened. Upon awakening, cages containing groups of two to three mice were transferred to either 24 or 34°C infant isolettes for the remainder of the experiment. In conscious unrestrained mice, such exposures maintain core temperatures of 36.5–37°C and 39.5–40°C, respectively (20). Room temperature was 23–24°C. To study the effect of FRH on survival in antibiotic-treated pneumonia, mice were inoculated i.t with 250 CFU K. pneumoniae and returned to 24°C cages for 48 h. Then half the mice were switched to 34.5°C isolettes (FRH) and treatment was initiated with 28 mg/kg ceftriaxone i.p per day.
Lung lavage and processing
At selected times, groups of mice were anesthetized by 10–30 s exposure to isoflurane, euthanized by cervical dislocation, and lung lavage was performed in situ through an 18-gauge blunt-end needle secured in the trachea, using 1 ml of PBS instilled and withdrawn twice, followed by instillation and recovery of a second 1 ml of PBS. The two aliquots of lung lavage were pooled, cells were collected by centrifugation at 1000 x g for 3 min, and cell-free supernatants were stored at –80°C for analysis of total protein and cytokine concentrations. Total cell counts were performed manually using a hemacytometer and differential cell counts of DiffQuick-stained cytopreparations (Baxter Scientific Products) were performed by two blinded observers (P. Rice and J. D. Hasday) using morphologic criteria.
Measurement of bacterial load
The lungs were weighed and homogenized in 1 ml of 0.9% NaCl between sterile frosted glass slides. The solid tissue was allowed to sediment for 10 min at room temperature. Homogenate supernatants and blood were serially diluted and 10-μl aliquots of each were plated on MacKonkey agar and incubated at 37°C overnight, and the number of colonies was enumerated and expressed per gram wet weight of lung tissue.
Measurement of cytokine and total protein concentration
Mouse TNF-, IFN-, IL-1, GM-CSF, MIP-2, and KC and human IL-8 were measured in the University of Maryland Cytokine Core Laboratory using standard two-Ab ELISA with commercial Ab pairs and recombinant standards (TNF- and IL-6 from Endogen; IFN-, GM-CSF, MIP-2, and KC from R&D Systems; and IL-1 from Genzyme) and IL-8 from BioSource International. Polystyrene plates (Maxisorb; Nunc) were coated with capture Ab in PBS overnight at 25°C. The plates were washed four times with 50 mM Tris, 0.2% Tween 20 (pH 7.2), and then blocked for 90 min at 25°C with assay buffer (PBS containing 4% BSA and 0.01% Thimerosal, pH 7.2). The plates were washed and 50 μl of assay buffer was added to each along with 50 μl of sample or standard prepared in assay buffer and incubated at 37°C for 2 h. After washing, strepavidin-peroxidase polymer in casein buffer (Research Diagnostics; Mount Pleasant, NJ) was added and incubated at 25°C for 30 min. The plate was washed and 100 μl of substrate (tetramethylbenzidine; DakoCytomation) was added and incubated for 20–30 min. The reaction was stopped with 100 μl of 2 N HCl, and the OD450 (minus OD650) was read on a microplate reader (Molecular Devices). The data was analyzed using a computer program (SoftPro; Molecular Devices). The TNF-, IL-6, IFN-, IL-1, GM-CSF, MIP-2, KC, and IL-8 assays had lower detection limits of 8, 3, 3.9, 1.5, 3, 61, 4, and 4 pg/ml, respectively. Total protein concentrations in supernatants were measured with a commercial reagent (Bio-Rad) with a standard curve constructed using BSA (Sigma-Aldrich).
Histologic analysis
After euthansasia, the anterior chest wall was removed; the trachea was cannulated with an 18-gauge blunt needle; the lungs were inflated in situ with 0.7 ml of 50% (v/v in PBS) OCT medium; the lungs and mediastinum were removed en bloc and embedded in neat OCT, and 8-μm cryosections were acetone-fixed for 3 min at –20°C and stored at –80°C for H&E staining and for immunostaining. A veterinary pathologist (T. O’Neil) evaluated the sections in a blinded fashion for morphologic signs of injury and inflammation. The total area of each lung section with inflammation was measured using a computer-assisted morphometric program (Image-Pro Plus; MediaCybernetics). In the HSF-1-null and control mice exposed to LPS and FRH, the extent of bronchiolar epithelial injury was graded by a blinded observer (J. D. Hasday) using a 0–4 scale where 0 was normal histologic appearance, 1 was loss of cilia, 2 was mild to moderate epithelial cell flattening, 3 was extensive epithelial cell flattening and nuclear degradation, and 4 was denuding of the epithelium with exposure of the underlying basement membrane. For immunohistologic analysis, slides were washed in PBS at 25°C for 20 min and endogenous peroxidase was inactivated by incubating for 10 min with 30% hydrogen peroxide in methanol at 25°C. Nonspecific signal was reduced by sequentially blocking with 5% serum (v/v) in PBS for 30 min, and then with a commercial avidin/biotin blocking kit (Vector Laboratories) according to the manufacturer’s instructions. The blocked slides were sequentially incubated with primary Ab for 1h, a secondary Ab (where indicated) for 30 min, a commercial avidin/biotin peroxidase detection system (Vector Laboratories), and 1 mg/ml diaminobenzidine (Sigma-Aldrich) in 0.02% hydrogen peroxide in PBS. All blocking and subsequent incubations were at 25°C. The blocking serum and primary and secondary Abs used for each cell type were as follows: for macrophages, mouse serum, 1 μg/ml rat anti-mouse Mac-3, and 2.5 μg/ml biotinylated goat anti-rat IgG1/2a HRP (both from BD Pharmingen). For PMN, bovine serum and the 2.5 μg/ml biotinylated rat anti-mouse Gr-1 (BD Pharmingen). For heat shock protein (HSP)72, 5% rabbit serum, 4 μg/ml goat anti-mouse HSP72 (Santa Cruz Biotechnology #SC-1060), and 1/200 dilution of rabbit anti-goat HRP (Sigma-Aldrich #A 5420). Immunostained slides were counterstained with hematoxylin.
Determination of myeloperoxidase activity in tissue
Following euthanasia, the pulmonary circulation was cleared with 10 ml of PBS injected through the left ventricle and drained through an incision in the right atrium. The lungs were dissected free of the trachea and mediastinal contents and homogenized in 1 ml of 50 mM potassium phosphate buffer, pH 6, containing 5 g/L hexadecyltrimethylammonium bromide, passed through a 21-gauge needle three times, sonicated, then subjected to three freeze-thaw cycles. Myeloperoxidase was measured in a 150-μl reaction containing 0.167 mg/ml o-dianisidine (Sigma-Aldrich) and 0.0005% hydrogen peroxide, 50 mM potassium phosphate buffer, pH 6, and 5 μl of sample. Following incubation at room temperature, the OD450 was measured and the myeloperoxidase activity was determined from a standard curve (Sigma-Aldrich). Total protein concentration was measured using Coomassie blue-based assay (Bio-Rad) and a BSA standard curve. Myeloperoxidase activity was standardized to protein concentration.
Epithelial cell culture models
A549 cells were obtained from the American Type Culture Collection and 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 (Invitrogen Life Technologies), pH 7.3 (CRPMI) and containing 10% defined FBS (HyClone) at 37°C in 5% CO2-enriched air. To measure IL-8 secretion, 1 x 105 A549 cells were plated in 1 ml of RPMI 1640/10% FBS in 24-well culture plates, and monolayers were established overnight. The culture medium was changed and, following a 30-min preincubation at the indicated temperature, 2 ng/ml recombinant human TNF- (R&D Systems) or 2 ng/ml IL-1 was added to the culture medium and incubated for 6 h at 37 or 39.5°C or at 42°C for the initial 2 h followed by 4 h at 37°C. IL-8 levels in the 6-h culture supernatants were measured by ELISA.
Immunoblotting
Cell extracts were prepared in RIPA buffer containing 10 mM Tris-HCl (pH 8.0), 150 mM NaCl, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, protease inhibitors (inhibitor cocktail; Boehringer Mannheim), 50 mM NaF, 20 mM -glycerophosphate, 1 mM sodium vanadate, 2 nM calyculin A, and 10 nM okadaic acid. For immunoblotting, 20 μg of total protein was separated by 7.5% SDS-PAGE; electrostatically transferred to polyvinylidene difluoride membrane (Stratagene); blocked with 5% nonfat dry milk in 25 mM Tris (pH 7.4), 0.5 M NaCl/0.05% (v/v) Tween 20 for 30 min at room temperature; and incubated with 1/10,000 dilution of anti-HSP72 Ab (Santa Cruz Biotechnology) in blocking buffer for 1 h. Bands were detected using 1/5000 dilution of goat anti-rabbit IgG HRP conjugate (Bio-Rad) in blocking buffer for 30 min at room temperature, developed with a chemiluminescence detection system (Renaissance; New England Nuclear), and exposed to x-ray film.
Data analysis
Normal data are presented as mean ± SEs and differences among more than two groups were analyzed by applying a post hoc Fisher protected least-squares difference to a one-way ANOVA and differences between two groups was analyzed by Student’s t test. Survival was analyzed with a Gehan-Wilcoxon test of a Kaplan-Meier plot. All experiments were performed at least two times and tested for reproducibility by two-way ANOVA before combining. The epithelial injury scores in the experiment with HSF-1-null mice and the bacterial loads in necropsied mice were not normal and were analyzed by the Mann-Whitney U test.
Results
Effect of FRH on survival in experimental K. pneumoniae pneumonia
We previously found that housing mice with experimental K. pneumoniae peritonitis at 34.5°C rather than 25°C increased core temperature from 36.5–37°C to 39.5–40°C and improved survival (20). To determine whether mice challenged with bacterial agonists via the airway route have a similar relationship between ambient and core temperatures, we telemetrically monitored core temperature in mice that were housed at either 25°C (euthermic) or 34.5°C (FRH) following i.t. challenge with 50 μg of LPS or inoculated with 750 CFU K. pneumoniae Caroli strain (Fig. 1A). In the LPS-challenged mice, exposure to FRH increased core temperature from 37.2 ± 0.2°C to 39.6 ± 0.2°C and in the i.t. K. pneumoniae-inoculated mice from 36.6 ± 0.10°C to 39.0 ± 0.09°C. To determine whether FRH improves survival in experimental pneumonia, groups of 10 mice were briefly anesthetized, i.t. inoculated with 250 CFU K. pneumoniae Caroli strain, and, upon waking, were placed in either 25 or 34.5°C cages and monitored for survival (Fig. 1B). In contrast with our previous observations that FRH improves survival in a K. pneumoniae peritonitis model (20), FRH failed to improve, and tended to worsen survival in mice with experimental K. pneumoniae pneumonia (p = 0.25). We previously demonstrated that the improved survival conferred by FRH in mice with experimental K. pneumoniae peritonitis was associated with a profound reduction in i.p. and systemic bacterial burden (20). FRH exerted similar effects on intrapulmonary bacterial burden in experimental K. pneumoniae pneumonia (Fig. 1C), reducing the median 72-h-postinoculation pulmonary bacterial burden 400-fold (p < 0.02). Little bacterial dissemination occurred in either euthermic (83 ± 75 CFU per milligram of spleen tissue) or FRH (0.9 ± 0.9 CFU per milligram of spleen tissue by 72 h postinoculation. Furthermore, necropsy analysis revealed that the median pulmonary bacterial burden at time of death was 85-fold lower in the warmer animals (4.68 x 104 vs 4.0 x 106 CFU per milligram of lung tissue, six mice per group; p = 0.016), suggesting that mechanisms other than proliferating bacteria contributed to death in the warmer animals.
FIGURE 1. Effect of FRH on survival and pulmonary pathogen burden in experimental K. pneumoniae pneumonia. A, Groups of four mice were implanted with i.p. thermistors and 1 wk later, were challenged with 50 μg of LPS i.t., (squares) or 750 CFU K. pneumoniae Caroli (circles) and placed at 24°C (euthermic; open symbols) or 34°C (FRH; closed symbols) and body temperature was measured telemetrically. B, Groups of 10 mice were inoculated with 250 CFU Klebsiella pneumoniae i.t., then housed at 24°C or 34°C, and survival was monitored. Duplicate experiments were performed. C, Groups of six euthermic and FRH-exposed mice were sequentially euthanized, and bacterial load in lung tissue was quantified. Mean ± SE. *, p < 0.05 vs euthermic mice.
To further analyze the cause of lethality in FRH-treated mice, we developed an antibiotic-treated K. pneumoniae pneumonia model. This model permits us to examine mechanisms of lethality in the absence of continued bacterial growth. Euthermic mice were i.t. inoculated with 250 CFU K. pneumoniae Caroli strain and, 48 h later, treatment with 28 mg/kg ceftriaxone i.p. per day was initiated. By day 5 of antibiotic treatment, proliferating bacteria were eliminated in lung tissue of euthermic animals (Fig. 2A). However, when mice were cotreated with FRH beginning simultaneously with antibiotic treatment, survival was reduced from 100% in the euthermic mice to 50% in the hyperthermic mice (Fig. 2B). Four of five deaths in the FRH group occurred within 2 days of initiating FRH exposure. At necropsy, the lungs in the FRH-treated mice were edematous and hyperemic. However, proliferating bacteria were undetectable in the lungs of three of five dead FRH-exposed mice, indicating that FRH did not antagonize and may have synergized with the bactericidal activity of the antibiotic. Furthermore, these observations suggest that despite the successful elimination of pathogens in experimental pneumonia, antibiotics failed to improve survival in FRH-treated mice.
FIGURE 2. Effect of FRH on survival in mice with antibiotic-treated pneumonia. A, Mice were inoculated with 250 CFU i.t. K. pneumoniae as described in Fig. 1 and, 2 days later, began treatment with 28 mg of ceftriaxone per killogram i.p. daily or sham-treated with i.p. saline. All mice remained euthermic throughout the experiment. Mean ± SE. *, p < 0.05 vs sham-treated mice. B, Groups of mice were inoculated with 250 CFU i.t. K. pneumoniae and, coincident with initiation of ceftriaxone therapy, mice were either switched to a 34°C cage (FRH) or maintained at 24°C (euthermic) and survival was monitored.
Coexposure to FRH and i.t. LPS causes lethal lung injury
To further elucidate the mechanisms of lung injury and death in mice coexposed to intrapulmonary bacterial pathogens and FRH, we developed an i.t. LPS-challenge model. Such a challenge with 50 μg of LPS was nonlethal in euthermic mice, but caused 50% mortality when administered to mice that were coexposed to FRH, with all of the deaths occurring within 3 days of i.t. LPS challenge (Fig. 3A). The increased mortality in the FRH-treated mice was associated with earlier and greater accumulation of protein in bronchoalveolar lavage fluid (BALF) (Fig. 3B), indicating acceleration and augmentation of LPS-induced pulmonary vascular endothelial injury in the warmer mice. By 48 h post-LPS challenge, BALF protein concentrations were 1.5 times greater in the warmer mice (8.5 ± 0.9 vs 5.6 ± 0.8 μg/ml; p < 0.02), indicating greater loss of pulmonary vascular endothelial barrier in the warmer mice.
FIGURE 3. FRH increases pulmonary vascular injury and reduces survival in mice challenged with i.t. LPS. Mice were challenged with 50 μg of LPS i.t. and housed at 24°C (euthermic) or 34°C (hyperthermic) and survival was measured (A). Groups of six mice were euthanized at each time point and pulmonary vascular leak was analyzed by measuring total BALF protein concentration (B). Mean ± SE. *, p < 0.05 vs euthermic mice.
Interstitial and alveolar inflammation was first evident by 12 h after LPS instillation in the euthermic mice. Subsequent progression of inflammation was accompanied by type II pneumocyte hypertrophy and mucous accumulation. FRH-exposed mice displayed qualitatively similar, but more extensive changes in lung histology compared with euthermic mice at each time point (Fig. 4, A–C). Interestingly, by 48 h after LPS challenge, widespread bronchiolar epithelial necrosis occurred in the FRH-exposed mice, but not in the euthermic animals (Fig. 5).
FIGURE 4. FRH increases the extent of lung consolidation in i.t. LPS-challenged mice. Mice were challenged with 50 μg of LPS i.t. then placed at 24°C or 34°C. Two mice per group were euthanized at each time point, the lungs were inflated with 0.7 ml of 50% (v/v) OCT, and 5-μm cryosections were stained with H&E and analyzed for area of consolidation using morphometric analysis software. Representative 20x photomicrographs of euthermic (A) and hyperthermic (B) mice 48 h after LPS challenge and means of all data (C) are shown.
FIGURE 5. FRH and i.t. LPS challenge causes bronchiolar epithelial necrosis. Mice were challenged with 50 μg of LPS i.t. then placed at 24°C (euthermic; B–C) or 34°C (hyperthermic; D–F). Two mice per group were euthanized per time point, the lungs were inflated with 0.7 ml of 50% (v/v) OCT, and 5-μm cryosections were stained with H&E. Representative photomicrographs (x200) of an untreated control (A), euthermic mice 36 h (B) and 48 h (C) after LPS challenge, and hyperthermic mice 24 h (D), 36 h (E), and 48 h (F) after LPS challenge. G, H, and I show high power views of A, C, and F, respectively. Note the loss of cilia and loss of distinct nuclei (arrows) in I.
Coexposure to FRH augments PMN accumulation within the LPS-challenged mouse lung
Exposing LPS-challenged mice to FRH augmented the accumulation of PMN in BALF 3-fold 48 h after i.t. LPS challenge from 1.66 ± 0.49 x 106 cells in the euthermic group to 4.1 ± 0.55 x 106 cells in the FRH group (Fig. 6A; p = 0.005). Microscopic analysis of lung tissue immunostained for the PMN marker GR-1 demonstrated PMN accumulation within inflammatory exudates of both groups of LPS-challenged mice; however, these exudates were more extensive in the FRH-exposed than the euthermic mice (Fig. 6, B and C). By 24 h after LPS challenge, myeloperoxidase activity in lung tissue was 2-fold higher in the FRH than euthermic mice (2 ± 0.23 vs 1.07 ± 0.24 myeloperoxidase units per microgram of protein; p < 0.02), providing further evidence of augmented pulmonary PMN accumulation in the warmer mice.
FIGURE 6. FRH augments pulmonary accumulation of PMN after i.t. LPS. Mice were challenged with 50 μg of LPS i.t. then placed at 24°C (euthermic) or 34°C (hyperthermic). Mice were euthanized at the indicated times; BALF was collected and analyzed for total PMN number (A). Mean ± SE; n = 6. *, p < 0.05 vs euthermic. Cryosections from hyperthemic (B) and euthermic (C) mice 48 h after i.t. LPS challenge were immunostained for the PMN marker, GR-1.
We previously demonstrated that FRH increased both pulmonary PMN burden and expression of the ELR+ CXC chemokines, KC, LIX/CXCL5, and MIP-2, and G-CSF in mice exposed to lethal hyperoxia (24). To determine whether FRH might also augment expression of these cytokines during Gram-negative bacterial pneumonia, we sequentially analyzed the cytokine composition of BALF from euthermic and FRH-exposed mice after i.t. LPS challenge (Fig. 7). BALF GM-CSF concentrations were 10-fold greater (253.3 ± 19.2 pg/ml vs 25.5 ± 22.2 pg/ml) in the FRH-exposed mice compared with euthermic animals 6 h after LPS challenge and IL-1 levels were increased 4.8-fold (382 ± 44.2 vs 78.9 ± 47.3 pg/ml) 24 h after LPS challenge. The kinetics of MIP-2 and KC expression were similar. Accumulation of both chemokines in BALF was evident 6 h after LPS challenge and reached comparable levels in the euthermic and hyperthermic mice, but a second peak of expression for each of these cytokines at 24 h post-LPS challenge only occurred in the hyperthermic mice. KC levels were 4.8-fold higher (1366 ± 239 vs 284 ± 106 pg/ml) and MIP-2 levels 2.3-fold higher (417 ± 64 vs 181 ± 67 pg/ml) in the FRH-exposed mice 24 h after LPS challenge compared with euthermic controls at the same time point. The BALF concentrations of TNF- and IFN- were similarly elevated in both groups (data not shown).
FIGURE 7. FRH augments BALF cytokine levels in i.t. LPS-challenged mice. Mice were challenged with 50 μg of LPS i.t., exposed to 24°C (euthermic) or 34°C (hyperthermic) ambient temperature. Groups of six mice were euthanized at each time point and lung lavage was analyzed for cytokine expression by ELISA. Mean ± SE. *, p < 0.05 vs euthermic animals.
Heat shock augments IL-8 secretion by TNF-- and IL-1-stimulated human lung epithelial cells
To determine whether FRH directly augments generation of ELR+ CXC chemokines by lung epithelium we analyzed the effect of in vitro FRH exposure on IL-8 expression by the lung alveolar epithelial-like cell line, A549 (Fig. 8, A and B). Because LPS did not directly activate IL-8 release in A549 cells (data not shown), we stimulated A549 cells with TNF- (Fig. 8A) or IL-1 (Fig. 8B), each of which is an early LPS-induced mediator which was elevated in lung lavage after i.t. LPS challenge. A549 cells released comparable levels of IL-8 at 37 and 39.5°C following stimulation with TNF- (1057 ± 35.7 pg per 106 cells at 37°C vs 862 ± 42.9 pg per 106 cells at 39.5°C) and IL-1 (2700 ± 209 vs 2414 ± 90 pg per 106 cells). In contrast, exposing A549 cells to 42°C for 2 h followed by 4 h incubation at 37°C augmented IL-8 expression 2-fold (2061 ± 120 pg per 106 cells) in TNF--stimulated cells and 2.3-fold (6513 ± 693 pg per 106 cells) in IL-1-stimulated cells compared with cells cultured at 37°C. Although exposing A549 cells to 42°C for 2 h stimulated expression of the inducible HSP, HSP72, exposure to 39.5°C for 6 h failed to activate a measurable increase in HSP72 protein levels (Fig. 8C), suggesting that heat shock may contribute to IL-8 expression.
FIGURE 8. Heat-shock augments TNF-- and IL-1-induced IL-8 expression in A549 cell culture. A549 cells (2 x 105 cells per milliliter of RPMI 1640/10% FBS) were incubated for 6 h with 10 ng/ml recombinant human TNF- (A) or 10 ng/ml IL-1 (B) at 37 or 39.5°C, or were exposed to 42°C for 2 h and switched to 37°C for 4 h (HS), and IL-8 secretion was measured by ELISA. *, p < 0.05 vs 37°C and vs 39.5°C. C, A549 cells were incubated for 4 h at 37°C (lane 1) or 39.5°C (lane 2) or were heated to 42°C for 2 h, then incubated for an additional 2 h at 37°C (HS; lane 3) and lysates were immunoblotted for HSP72 (arrow). A nonspecific band appeared at 50 kDa (arrowhead).
Although direct exposure of A549 cells to 39.5°C in vitro failed to activate a detectable heat shock response, Ostberg et al. (27) previously reported that exposing intact mice to a core temperature of 39.5°C for 6 h activates heat shock response in several organs, including the lungs. To determine whether FRH activates heat shock in the i.t. LPS-challenged mouse lung and identify the cellular expression pattern, we analyzed lung cryosections that were immunostained for anti-HSP72 (Fig. 9). Coexposure to FRH and LPS for 24 h induced expression of HSP72 primarily within the bronchiolar and bronchial epithelium (Fig. 9A, see arrows). In contrast, only minimal HSP72 expression was evident in the lungs of LPS-challenged, euthermic mice.
FIGURE 9. FRH induces expression of HSP72 in bronchiolar epithelium. Mice were challenged with 50 μg of LPS i.t. then placed at 24°C (euthermic) or 34°C (hyperthermic) for 24 h. Cryosections from hyperthermic (A) and euthermic (B) mice 24 h after i.t. LPS challenge or untreated control mice (C) were immunostained for HSP72. Arrows in A indicate areas of increased bronchiolar epithelial HSP72 staining.
To determine whether the heat-shock response activated in mice exposed to FRH contributes to the pattern of injury induced by coexposure to FRH and i.t. LPS, we compared the histological changes within the lung of HSF-1-null mice and their heterozygous littermates 24 h after exposure to 50 μg of LPS i.t. and FRH (Fig. 10). The heterozygous mice showed an injury pattern similar to that seen in the CD-1 mice exposed to LPS and FRH, with diffuse bronchiolar epithelial cell necrosis. In contrast, the bronchiolar epithelium remained well-preserved in the HSF-1-null mice despite exposure to the same combination of LPS and FRH.
FIGURE 10. HSF-1-null mice are resistant to bronchiolar epithelial necrosis after exposure to LPS and FRH. HSF-1-null mice (A) and their heterozygous littermates (B) were challenged with 50 μg of LPS i.t. then placed at 34°C. Four mice per group were euthanized 24 h later, the lungs were inflated with 0.8 ml of 50% (v/v) OCT, and 5-μm cryosections were stained with H&E. Representative x100 photomicrographs of HSF-1-null (A) and heterozygous (B) mice are shown. C, The summary data of arbitrary bronchiolar epithelial injury scores of the two groups (mean ± SE; p = 0.011).
Discussion
We have previously shown that increasing core temperature from basal to febrile levels improves survival in K. pneumoniae peritonitis by accelerating pathogen clearance (20). However, death still occurred in some of the warmer mice despite reduced pathogen loads, suggesting that the host response rather than the pathogen itself may contribute to death in the warmer mice. Nonetheless, the improved survival in the FRH-treated mice demonstrates that in the peritonitis model, the beneficial effects of FRH on pathogen clearance outweigh its potential harmful effects. In the present study, we extended these observations by analyzing the effect of FRH on immune responses to a bilateral multilobar Gram-negative bacterial pneumonia, an antibiotic-treated pneumonia, and i.t. LPS. We found that FRH failed to improve survival in the K. pneumoniae pneumonia model despite affecting a profound reduction in pulmonary bacterial burden. We showed that antibiotic-treated pneumonia and i.t. LPS challenge, each of which was nonlethal in euthermic mice, caused 50 and 45% mortality, respectively, when accompanied by FRH. In the intratrachal LPS model, exposure to FRH increased pulmonary PMN accumulation, pulmonary vascular endothelial leak, and bronchiolar epithelial injury. Hyperthermia also augmented IL-1, MIP-2, KC, and GM-CSF secretion into the bronchoalveolar compartment. These data suggest that, in the setting of diffuse bacterial pneumonia, FRH increases generation of proinflammatory cytokines, increases recruitment of PMN, augments immune-mediated pulmonary endothelial and epithelial injury, and increases mortality.
Our temperature-controlled mouse model was designed to isolate the effects of febrile temperature itself from the processes that generate fever. The effects of FRH on pulmonary inflammation and epithelial injury were not due to direct airway warming, as the temperature of the inhaled air (34°C) was below normal core temperature (37°C). The use of anesthesia during the inoculation procedure avoided the effects of physical stress and provided more precise control of the amount of inoculum entering the lungs. Although anesthetic agents may influence the acute-phase response, three lines of evidence suggest that the anesthetics used in this study did not interfere with the studied immune parameters. First, previous studies in our laboratory found that use of tribromoethanol had no effect on LPS-induced cytokine expression in euthermic mice, nor did anesthesia modify the effects of hyperthermia on LPS-induced cytokine expression (28). Second, in the present study, the euthermic and hyperthemic mice were treated with the same anesthesia protocol. Third, no toxic effects of anesthesiaadministration were found in temperature-controlled mice mock challenged with sterile saline i.t.
We previously reported that exposing mice to FRH activates expression of circulating G-CSF, expansion of the circulating PMN pool, and, in mice exposed to hyperoxia for 24 h, enhances bronchoalveolar expression of the ELR+ CXC chemokines, KC, LIX, and MIP-2 (24). In the present study, i.t. LPS induced early expression of KC and MIP-2 in BALF evident at 6 h following LPS challenge, which was not modified by coexposure to FRH. However, a second peak of BALF KC and MIP-2 occurred 24 h after LPS challenge, but only in the mice exposed to FRH. Exposure to FRH stimulated a similar increase in BALF IL-1 levels 24 h post-LPS challenge that temporally correlated with the augmented chemokine expression in the warmer mice. In contrast with the effects of FRH on chemokine expression, coexposure to FRH augmented early GM-CSF expression evident at 6 h post-LPS challenge by 10-fold, but did not stimulate subsequent GM-CSF expression. Furthermore, while the effects of FRH on KC and MIP-2 expression in LPS-challenged mice in the present study were similar to those previously found in hyperoxia-exposed mice (24), FRH failed to augment BALF GM-CSF or IL-1 levels in the hyperoxia mouse model (24). Although the proximal signaling events that mediate pulmonary oxygen toxicity are not well understood, the susceptibility of TLR4-defective, LPS-resistant C3H/HeJ mice to hyperoxia-induced lung inflammation and injury (29) suggests that the lung injury induced by LPS and by hyperoxia are triggered by distinct signaling events. Collectively, these data suggest that FRH exerts specific effects on a subset of cytokine genes and that the pattern of gene expression is determined in part by the proximal injurious agent. Interestingly, the pattern of augmented cytokine generation in these models suggests a coordinated pattern of biologic actions through which fever may augment PMN activation and delivery to sites of infection and injury. We previously found that the augmented PMN accumulation in the lungs of mice exposed to hyperoxia was completely blocked by immunoblockade of the common receptor for CXC chemokines (24). However, in the current study, PMN content continued to increase between 24 and 48 h after LPS challenge in the FRH-exposed mice despite a peak and subsequent decline in BALF levels of MIP-2 and KC over this same time. These data suggest that additional chemotactic factors (e.g., complement protein C5a fragment or leukotriene B4) may contribute to the late PMN accumulation in the warmer mice. Furthermore, the observation by Bernheim et al. (18) that exposure to FRH causes a similar accumulation of granulocytes at sites of Gram-negative infection in a more primitive animal, the desert iguana Diposuaurus dorsalis, suggests that these processes may be phylogenetically old and evolutionarily conserved.
We showed that exposure to core temperature of 39.5°C was sufficient to stimulate an increase in HSP72 immunostaining in the LPS-challenged mouse lung. The anti-HSP72 Ab, which specifically recognizes the heat-shock-inducible HSP70 family member, failed to immunostain lung tissue from euthermic mice. These results are consistent with those of Ostberg et al. (27), in which they showed that exposing mice to 39.5–40°C for 6 h induced expression of HSP72 and HSP110 in several organs, including lung. Furthermore, the thermal threshold for heat shock may be reduced further by inflammatory mediators that may be present in the infected or injured lung, including arachidonic acid (30), type I IFNs (31), and oxidants (32). Collectively, these data support the concept that heat shock activation may occur during febrile illnesses.
We have previously shown that FRH (39.5°C) causes only modest augmentation in TNF--induced expression of the human ELR+ CXC chemokine, IL-8, and GM-CSF in cultured human pulmonary artery endothelial cells (33). In the present study, in vitro exposure to FRH (39.5°C) failed to augment detectable IL-8 or GM-CSF expression in A549 cultured epithelial cells stimulated by either TNF- or IL-1. However, heating the same cells to 42°C for 2 h augmented TNF-- and IL-1-induced expression of IL-8 and GM-CSF and induced a heat shock response as indicated by the appearance of HSP72 protein. TNF- and IL-1 use distinct proximal signaling pathways that converge on NF-B, a transcription factor critical for IL-8 and GM-CSF gene activation. However, several studies have demonstrated that heat shock blocks NF-B activation (34). Yoo et al. (35) reported that pre-exposure to heat shock reduced NF-B activation and IL-8 secretion in BEAS-2B and A549 cultured respiratory epithelial cells by inhibiting IKK and stabilizing IB. In contrast with the coincident exposure of A549 cells to heat shock and proinflammatory stimuli analyzed in the present study, the studies showing an inhibitory effect of heat shock on NF-B activation used a protocol in which the inflammatory stimulus was added after heat shock had been induced. On the other hand, Bowman et al. (36) showed that exposing cultured human keratinocytes to 54°C for 4 s activated expression of both HSP72 and IL-8. These studies are limited to showing an association between heat shock activation and expression of ELR+ CXC chemokines. Additional studies establishing a regulatory role for heat shock pathways in activation of the ELR+ CXC chemokine gene family are ongoing in our laboratory.
It is notable that in the present study, exposing mice to FRH augmented BALF IL-1 levels 24 h after LPS challenge. In contrast, we and others have shown that direct exposure of macrophages in vitro to FRH (39.5°C) (37) or heat shock (42°C) temperature (38) attenuates IL-1 expression, in part by inhibiting C/EBP-mediated transcription (38). Possible explanations for the discrepancy between the in vitro and in vivo effects of heating include: 1) heat shock exerts different effects on IL-1 transcription in lung macrophages in vivo and monocytes and macrophages in vitro; 2) the augmentation of IL-1 secretion in vivo is mediated through posttranscriptional effects of heat shock; and 3) that cells other than macrophages secrete the excess IL-1 in the warmer mice. The temporal correlation between the rise in BALF IL-1 levels and PMN influx and the known capacity of PMN to express IL-1 (39) implicates this cell as one potential source of excess IL-1 in the FRH-exposed mice. However, definitive identification of the cellular source of excess IL-1 in the FRH-exposed, LPS-challenged lung awaits the result of ongoing studies in our laboratory.
Although several studies by other groups have shown that hyperthermia protects against pulmonary injury (40, 41, 42, 43), the heating protocols used in these studies achieved higher temperatures, 40.5–42°C, for shorter periods, 15–20 min, and allowed for recovery of up to 18 h before exposure to injurious agents. In contrast with the heat shock preconditioning protocols used in these studies, the warming protocol used in the present study was designed to model fever associated with naturally occurring infections. Using a similar coexposure protocol, Suzuki et al. (44) recently reported that warming rabbits 2°C above their basal core temperature augmented lung injury induced by high tidal volume mechanical ventilation in vivo and in isolated lung preparations perfused with 2% blood ex vivo. However, because lung injury in this model was evident after only 2 h of hyperthermia and mechanical ventilation and occurred in the excised lungs despite the absence of PMN in the perfusate, additional mechanisms may have contributed to lung injury in the warmer mice in this study. We previously demonstrated that FRH (39.5°C) augments TNF--induced paracellular leak in cultured human pulmonary artery endothelial cells in vitro (33), thereby demonstrating another potential effect of FRH that may contribute to lung injury.
Whereas, cytokine expression, PMN recruitment, and pulmonary vascular endothelial leak were stimulated by LPS challenge in both euthermic and hyperthermic mice, albeit to higher levels in the warmer mice, extensive bronchiolar epithelial necrosis occurred only in the hyperthermic mice. Expression of HSP72, an indicator of the heat shock response, was also evident in the bronchial and bronchiolar epithelium of the warmer mice. Although the temporal and spatial correlation between epithelial necrosis and HSP72 expression suggests a possible mechanistic link, we have not yet generated evidence that heat shock itself leads to bronchiolar epithelial necrosis in the FRH-exposed, LPS-challenged lung. Although most studies have concluded that induction of thermotolerance or overexpression of HSP72 blocks apoptosis (45, 46, 47), elicitation of the heat shock state preceded exposure to the apoptotic stimulus in these studies. In contrast, in febrile illnesses, HSPs are expressed during the inflammatory response. Furthermore, Buzzard et al. (47) reported that overexpression of HSP72 in mouse embryonic fibroblasts exerted anti-apoptotic action against some (e.g., TNF- and heat) but not all (e.g., ionizing radiation) apoptotic stimuli. Similarly, Gabai et al. (46) reported that HSP72 overexpression in human fibroblasts blocked early, Bid-dependent apoptosis, but not subsequent Bid-independent apoptosis. It is also possible that the bronchiolar epithelial necrosis that occurred in the warmer mice is not caused directly by the expression of HSP genes within the epithelium. These animals were not only subjected to a greater intrapulmonary PMN burden, but they also were exposed to higher levels of the PMN activating cytokines, GM-CSF, KC, MIP-2, and IL-1. Furthermore, FRH itself directly enhances PMN generation of reactive oxygen intermediates and NO responsiveness to LPS (48). In fact, we previously reported that immunoblockade of the common receptor for KC and MIP-2 in hyperoxia-exposed mice partially reversed the accelerated death caused by coexposure to FRH (24).
In summary, we have shown that the same exposure to FRH that was protective in K. pneumoniae peritonitis (20), failed to improve survival in pneumonia with the same pathogen. These results suggest that FRH exerts profound immunomodulating effects during infections, and that the consequences of fever are determined by a balance between accelerated pathogen clearance and augmented collateral tissue injury. The disparate effect of FRH on survival in bacterial peritonitis and pneumonia indicates that the ultimate effects of fever are determined by the site of infection or injury as well, and that the lung may be particularly susceptible to augmented tissue injury during fever. These results show that there is a need for future studies of fever management in well-defined patient populations.
Disclosures
The authors have no financial conflict of interest.
Acknowledgments
We thank Dr. Matthew Fenton for his thoughtful review of the manuscript.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 This work was supported by National Institutes of Health Grants HL69057, GM066855, and AI42117 and a Veterans Affairs Merit Review Award.
2 Address correspondence and reprint requests to Dr. Jeffrey D. Hasday, Baltimore Veterans Administration Medical Center, Room 3D122, 10 North Greene Street, Baltimore, MD 21201. E-mail address: jhasday{at}umaryland.edu
3 Abbreviations used in this paper: FRH, febrile-range hyperthermia; HSP, heat shock protein; i.t., intratracheal; PMN, polymorphonuclear cell; BALF, bronchoalveolar lavage fluid.
Received for publication August 20, 2004. Accepted for publication December 27, 2004.
References
Kluger, M. J., W. Kozac, C. A. Conn, L. R. Leon, D. Soszynski. 1996. The adaptive value of fever. P. A. Mackowiak, ed. Fever: Basic Mechanisms and Management 255. Raven Press, New York.
Doran, T. F., C. DeAngelis, R. A. Baumgardner, E. D. Mellits. 1989. Acetaminophen: more harm than good for chicken pox?. J. Pediatr. 114:1045.
Graham, N. M. H., C. J. Burrell, R. M. Douglas, P. Debelle, L. Davies. 1990. Adverse effects of aspirin, acetaminophen, and ibuprofen on immune function, viral shedding, and clinical status in rhinovirus-infected volunteers. J. Infect. Dis. 162:1277
Stanley, E. D., G. G. Jackson, C. Panusarn, M. Rubenis, V. Dirda. 1975. Increased viral shedding with aspirin treatment of rhinovirus infection. J. Am. Med. Assoc. 231:1248.
Mackowiak, P. A., S. S. Wasserman, M. M. Levine. 1992. An analysis of the quantitative relationship between oral temperature and severity of illness in experimental shigellosis. J. Infect. Dis. 166:1181.
Bryant, R. E., A. F. Hood, C. E. Hood, M. G. Koenig. 1971. Factors affecting mortality of gram-negative rod bacteremia. Arch. Intern. Med. 127:120.
Mackowiak, P. A., R. H. Browne, P. M. Southern, Jr, J. W. Smith. 1980. Polymicrobial sepsis: an analysis of 184 cases using log linear models. Am. J. Med. Sci. 280:73.
Weinstein, M. P., P. B. Iannini, C. W. Stratton, T. C. Eickhoff. 1978. Spontaneous bacterial peritonitis: a review of 28 cases with emphasis on improved survival and factors influencing prognosis. Am. J. Med. 64:592.
Hoefs, J. C., H. N. Canawati, F. L. Sapico, R. R. Hopkins, J. Weiner, J. Z. Montgomerie. 1982. Spontaneous bacterial peritonitis. Hepatology 2:399.
Ahkee, S., L. Srinath, J. Ramirez. 1997. Community-acquired pneumonia in the elderly: association of mortality with lack of fever and leukocytosis. South Med. J. 90:296
Hasday, J. D., K. D. Fairchild, C. Shanholtz. 2000. The role of fever in the infected hosts. Microbes Infect. 2:1891
Vaughn, L. K., H. A. Bernheim, M. J. Kluger. 1974. Fever in the lizard Dipsosaurus dorsalis. Nature 252:473.
Armstrong, C.. 1942. Some recent research in the field of neurotropic viruses with especial reference to lymphocytic choriomeningitis and herpes simplex. Milit. Surg. 91:129.
Bell, J. F., G. F. Moore. 1974. Effects of high ambient temperature on various stages of rabies virus infection in mice. Infect. Immun. 10:510.
Covert, J. B., W. W. Reynolds. 1977. Survival value of fever in fish. Nature 267:43
Bernheim, H. A., P. T. Bodel, P. W. Askenase, E. Atkins. 1978. Effects of fever on host defence mechanisms after infection in the lizard Dipsosaurus dorsalis. Br. J. Exp. Pathol. 59:76
Bernstein, I. L., H. S. Boushey, R. M. Cherniack, J. N. Fink, J. D. Fulmer, E. J. Goetzl, R. A. Goldstien, S. S. Hurd, J. S. Ram, L. M. Lichtenstein, et al 1985. Summary and recommendations of a workshop in the investigative use of fiberoptic bronchoscopy and bronchoalveolar lavage in asthmatics. Am. Rev. Respir. Dis. 132:180.
Jiang, Q., A. S. Cross, I. S. Singh, T. T. Chem, R. M. Viscardi, J. D. Hasday. 2000. Febrile core temperature is essential for optimal host defense in bacterial peritonitis. Infect. Immun. 68:1265.[Abstract/Free Full Text]
Kluger, M. J., D. H. Ringler, M. R. Anver. 1975. Fever and survival. Science 188:166.
Schmidt, J. R., A. F. Rasmussen, Jr. 1960. The influence of environmental temperature on the course of experimental herpes simplex infection. J. Infect. Dis. 107:356
Ellingson, H. V., P. F. Clark. 1942. The influence of artificial fever on mechanisms of resistance. J. Immunol. 43:65
.
Hasday, J., A. Garrison, I. Singh, T. Standiford, G. Ellis, S. Rao, J.-R. He, P. Rice, M. Frank, S. Goldblum, R. Viscardi. 2003. Febrile-range hyperthermia augments pulmonary neutrophil recruitment and amplifies pulmonary oxygen toxicity. Am. J. Pathol. 162:2005
Xiao, X., X. Zuo, A. A. Davis, D. R. McMillan, B. B. Curry, J. A. Richardson, I. J. Benjamin. 1999. HSF1 is required for extra-embryonic development, postnatal growth, and protection during inflammatory response in mice. EMBO J. 18:5943.
Keane-Myers, A. M., W. C. Gause, F. D. Finkelman, X. D. Xhou, M. Wills-Karp. 1998. Development of murine allergic asthma is dependent upon B7-2 costimulation. J. Immunol. 160:1036.
Ostberg, J. R., K. C. Kaplan, E. A. Repasky. 2002. Induction of stress proteins in a panel of mouse tissues by fever-range whole body hyperthermia. Int. J. Hyperthermia 18:552.
Jiang, Q., L. DeTolla, I. Kalvakolanu, B. Fitzgerald, J. D. Hasday. 1999. Fever upregulates expression of pyrogenic cytokines in endotoxin-challenged mice. Am. J. Physiol. 276:R1653.
Johnston, C. J., B. R. Stripp, B. Piedbeouf, T. W. Wright, G. W. Mango, C. K. Reed, J. N. Finkelstein. 1998. Inflammatory and epithelial responses in mouse strains that differ in sensitivity to hyperoxic injury. Exp. Lung Res. 24:189.
Jurivich, D. A., L. Sistonen, K. D. Sarge, R. I. Morimoto. 1994. Arachidonate is a potent modulator of human heat shock gene transcription. Proc. Natl. Acad. Sci. USA 91:2280.
Bowman, P. D., S. T. Schuschereba, D. F. Lawlor, G. R. Gilligan, J. R. Mata, D. R. DeBaere. 1997. Survival of human epidermal keratinocytes after short-duration high temperature: synthesis of HSP70 and IL-8. Am. J. Physiol. 272:C1988.
Fairchild, K. D., R. M. Viscardi, L. Hester, I. S. Singh, J. D. Hasday. 2000. Effects of hypothermia and hyperthermia on cytokine production by cultured human mononuclear phagocytes from adults and newborns. J. Interferon Cytokine Res. 20:1049
Gabai, V. L., K. Mabuchi, D. D. Mosser, M. Y. Sherman. 2002. Hsp72 and stress kinase c-jun N-terminal kinase regulate the bid-dependent pathway in tumor necrosis factor-induced apoptosis. Mol. Cell Biol. 22:3415.
Buzzard, K. A., A. J. Giaccia, M. Killender, R. L. Anderson. 1998. Heat shock protein 72 modulates pathways of stress-induced apoptosis. J. Biol. Chem. 273:17147
Rosenspire, A. J., A. L. Kindzelskii, H. R. Petty. 2002. Cutting edge:fever-associated temperatures enhance neutrophil responses to lipopolysaccharide: a potential mechanism involving cell metabolism. J. Immunol. 169:5396(Penelope Rice, Erica Mart)