Death receptors mediate the adverse effects of febrile-range hyperthermia on the outcome of lipopolysaccharide-induced lung injury

Abstract

We have shown that febrile-range hyperthermia enhances lung injury and mortality in mice exposed to inhaled LPS and is associated with increased TNF-α receptor activity, suppression of NF-κB activity in vitro, and increased apoptosis of alveolar epithelial cells in vivo. We hypothesized that hyperthermia enhances lung injury and mortality in vivo by a mechanism dependent on TNF receptor signaling. To test this, we exposed mice lacking the TNF-receptor family members TNFR1/R2 or Fas (TNFR1/R2(-/-) and lpr) to inhaled LPS with or without febrile-range hyperthermia. For comparison, we studied mice lacking IL-1 receptor activity (IL-1R(-/-)) to determine the role of inflammation on the effect of hyperthermia in vivo. TNFR1/R2(-/-) and lpr mice were protected from augmented alveolar permeability and mortality associated with hyperthermia, whereas IL-1R(-/-) mice were susceptible to augmented alveolar permeability but protected from mortality associated with hyperthermia. Hyperthermia decreased pulmonary concentrations of TNF-α and keratinocyte-derived chemokine after LPS in C57BL/6 mice and did not affect pulmonary inflammation but enhanced circulating markers of oxidative injury and nitric oxide metabolites. The data suggest that hyperthermia enhances lung injury by a mechanism that requires death receptor activity and is not directly associated with changes in inflammation mediated by hyperthermia. In addition, hyperthermia appears to enhance mortality by generating a systemic inflammatory response and not by a mechanism directly associated with respiratory failure. Finally, we observed that exposure to febrile-range hyperthermia converts a modest, survivable model of lung injury into a fatal syndrome associated with oxidative and nitrosative stress, similar to the systemic inflammatory response syndrome.

Figures

Fig. 1.

Exposure to ambient hyperthermia generates febrile-range core body temperature in sentinel animals. Exposure to ambient hyperthermia (35°C) generates febrile-range core body temperatures (39.5–40°C) in mice (n = 30) compared with animals that were maintained at room temperature (23°C) (n = 15) independent of LPS exposure.

Fig. 2.

Deficiencies in the activity of death receptors TNFR1/R2 or Fas, or in IL-1 signaling attenuate mortality associated with exposure to LPS and febrile-range hyperthermia. A–C: C57BL/6 wild-type (WT) or genetically abnormal mice in a C57BL/6 background were exposed to intratracheal LPS (50 μg) with or without febrile-range hyperthermia for up to 48 h. In all experiments, mice exposed to LPS and euthermia have 0% mortality (number exposed to LPS + euthermia for 48 h: C57BL/6 = 15, TNFR1/R2−/− = 3; lpr = 3; IL-1R−/− = 3). A: TNFR1/R2−/− mice exposed to LPS and hyperthermia (n = 12) have 0% mortality by 48 h compared with WT mice that, exposed to hyperthermia and LPS (n = 30), have 64% mortality. B: lpr mice exposed to LPS and hyperthermia (n = 12) have 17% mortality compared with 64% mortality in WT animals. C: IL-1R−/− mice exposed to LPS and hyperthermia (n = 12) have 0% mortality compared with 71% mortality in WT mice. **P < 0.0001 vs. WT hyperthermic animals.

Fig. 3.

A deficiency in the activity of death receptors TNFR1/R2 or Fas but not in IL-1 receptor activity attenuates hyperthermia-augmented lung injury. IgM (A) and total protein (B) concentrations in bronchoalveolar lavage (BAL) fluid collected at 48 h are increased in WT or IL-1R−/− animals exposed to LPS and hyperthermia compared with WT or IL-1R−/− animals exposed to LPS and euthermia or hyperthermia alone (data not shown). Hyperthermia does not affect IgM and total protein concentrations in TNFR1/R2−/− or lpr animals exposed to LPS. Total cell (C) and polymorphonuclear neutrophils (PMN) (D) concentrations in BAL fluid collected at 48 h were similar in all groups exposed to LPS and euthermia or hyperthermia with the exception of lpr mice, which had increased concentrations of cells and PMN in BAL fluid. Lung homogenate myeloperoxidase (MPO) activity (E) was unaffected by the presence of hyperthermia in all groups exposed to LPS for 48 h, *P < 0.05, §P < 0.001. Animal numbers in each experimental group are noted in Table 1.

Fig. 4.

A deficiency in the activity of the death receptors TNFR1/R2 or Fas, but not IL-1 receptor deficiency, protects from hyperthermia-augmented lung injury and apoptosis. Hematoxylin and eosin stains of paraffin-embedded lungs of wild-type (D–F) and IL-1R−/− animals (M–O) exposed to LPS and hyperthermia for 48 h show increased alveolar septal thickening and TUNEL-positive cells (indicated by arrows) compared with wild-type animals exposed to LPS and euthermia (A–C) or TNFR1/R2−/− (G–I) or lpr (J–L) animals exposed to LPS and hyperthermia. The histological examination of lung tissue sections of wild-type, lpr, and IL-1R−/− animals show similar PMN tissue recruitment, whereas TNFR1/R2−/− animals show significantly fewer PMNs.

Fig. 5.

A deficiency in TNFR1/R2 or Fas activity, but not in IL-1R activity, protects mice from apoptotic cell death after exposure to LPS and fever. Paraffin-embedded sections of lungs of mice treated with intratracheal LPS and euthermia or hyperthermia for 48 h were evaluated for apoptosis using the TUNEL method. The number of positive cells were quantified in a blinded manner from 10 randomly generated fields at ×100 magnification and compared. Compared with the lungs of C57BL/6 mice treated with LPS alone, the lungs of C57BL/6 animals treated with LPS and fever have significantly more TUNEL-positive cells. TNFR1/R2−/− or lpr animals treated with LPS and fever have significantly fewer TUNEL-positive cells than C57BL/6 animals treated with LPS and fever. The IL-1R−/− mice treated with LPS and fever had a similar number of TUNEL-positive cells in lung sections compared with C57BL/6 mice treated with LPS and fever, *P < 0.05.

Fig. 6.

Hyperthermia is associated with decreased concentrations of several important pulmonary cytokines after LPS-induced lung injury. Lung homogenates of wild-type animals exposed to LPS intratracheally (50 μg) and euthermia or hyperthermia for 24 or 48 h demonstrate that hyperthermia is associated with significantly lower concentrations of the cytokine TNF-α and the chemokine keratinocyte-derived chemokine (KC) in the lung at 48 h and a trend toward lower concentrations of granulocyte/macrophage colony-stimulating factor (GM-CSF) and IL-1β although these changes did not reach statistical significance. Hyperthermia significantly increases the concentration of macrophage inflammatory protein-2 (MIP-2) in lung homogenates of animals exposed to LPS and hyperthermia, *P < 0.05, **P < 0.01.

Fig. 7.

Hyperthermia increases plasma concentration of markers of oxidative stress and nitric oxide (NO) metabolites after LPS-induced lung injury. Plasma concentrations of malonyldialdehyde (MDA) (A) and nitric oxide metabolites (B) were measured in animals exposed to LPS and euthermia or hyperthermia for 24 or 48 h. Wild-type animals exposed to LPS and hyperthermia had significantly higher plasma concentrations of MDA and NO metabolites compared with wild-type animals exposed to LPS alone or hyperthermia alone. Wild-type animals exposed to LPS and hyperthermia also had significantly higher plasma concentrations of MDA and NO metabolites compared with animals lacking TNFR1/R2, Fas, or IL-1R activity exposed to LPS and hyperthermia, *P < 0.05, **P < 0.01, §P < 0.001 (ANOVA among all groups). Post hoc comparisons using the Bonferroni-Dunn test demonstrate that C57BL/6 animals treated with LPS and hyperthermia for 24 and 48 h had significantly higher MDA and NO metabolite concentrations than all other groups (P < 0.05).