Febrile-range hyperthermia augments pulmonary neutrophil recruitment and amplifies pulmonary oxygen toxicity

Abstract

Febrile-range hyperthermia (FRH) improves survival in experimental infections by accelerating pathogen clearance, but may also increase collateral tissue injury. We hypothesized that FRH would worsen the outcome of inflammation stimulated by a non-replicating agonist and tested this hypothesis in a murine model of pulmonary oxygen toxicity. Using a conscious, temperature-controlled mouse model, we showed that maintaining a core temperature at FRH (39 degrees C to 40 degrees C) rather than at euthermic levels (36.5 degrees C to 37 degrees C) during hyperoxia exposure accelerated lethal pulmonary vascular endothelial injury, reduced the inspired oxygen threshold for lethality, induced expression of granulocyte-colony stimulating factor, and expanded the circulating neutrophil pool. In these same mice, FRH augmented pulmonary expression of the ELR(+) CXC chemokines, KC and LPS-induced CXC chemokine, enhanced recruitment of neutrophils, and changed the histological pattern of lung injury to a neutrophilic interstitial pneumonitis. Immunoblockade of CXC receptor-2 abrogated neutrophil recruitment, reduced pulmonary vascular injury, and delayed death. These combined data demonstrate that FRH may enlist distinct mediators and effector cells to profoundly shift the host response to a defined injurious stimulus, in part by augmenting delivery of neutrophils to sites of inflammation, such as may occur in infections. In certain conditions, such as in the hyperoxic lung, this process may be deleterious.

Figures

Figure 1.

Effect of febrile range hyperthermia on survival time in mice exposed to hyperoxia. Mice were exposed to > 0.95 FIO2 (hyperthermic) or air (normoxic) and housed at either 24°C (euthermic) or 34.5°C (hyperthermic). A: Mean (±SE) core temperature was measured telemetrically in euthermic (squares; n = 4) and hyperthermic (circles; n = 4) mice. Open symbols are normoxic and closed symbols hyperoxic mice. * and † denote P < 0.05 versus euthermic-hyperoxic and euthermic-normoxic mice, respectively. B: Survival was determined in euthermic (n = 16) and hyperthermic (n = 16) mice.

Figure 2.

Effect of febrile range hyperthermia on lethal FIO2 threshold. Groups of 10 mice each were exposed to the indicated FIO2 while housed at either 24°C (squares) or 34°C to 34.5°C (circles) and were monitored for survival for 2 weeks.

Figure 3.

Effect of febrile range hyperthemia on hyperoxic lung injury. Mice were exposed to FIO2 > 0.95 and housed at either 24°C (euthermia) or 34.5°C (hyperthermia). A: BALF protein concentration: groups of eight mice were sacrificed after the indicated exposure time, lung lavage was performed, and total protein was measured. B: Wet:dry weight: groups of four mice were sacrificed after 36 hours, the lungs were excised, weighed before and after drying, and the wet:dry weight ratio was calculated. C: BALF IL-6 concentration: groups of eight mice were sacrificed after the indicated exposure time to FIO2 > 0.95 (hyperoxia) or FIO2 0.21 (normoxia), lung lavage was performed, and IL-6 concentration was measured by ELISA. Mean ± SE; * denotes P < 0.05 versus euthermic normoxic controls. ** denotes P < 0.05 versus euthermic mice at the same exposure time. Because hyperthermic mice did not survive beyond 60 hours of exposure to hyperthermia, analysis of later time points could not be done (ND).

Figure 4.

Influence of core temperature on the pattern of lung injury in mice exposed to hyperoxia. Representative hematoxalyn and eosin-stained sections from mice exposed to FIO2 > 0.95 while euthermic for 36 hours (A) or 108 hours (B) or while hyperthermic for 36 hours (C). Arrowhead in B indicates hyaline membranes; arrows in C indicate intraseptal PMN. Magnification, ×400.

Figure 5.

Effect of febrile range hyperthermia on inflammatory cell accumulation in hyperthermic lung. Euthermic and hyperthermic mice were exposed to FIO2 > 0.95. Groups of four mice were sacrificed at the indicated time, lung sections were immunostained for either PMN (anti-Gr-1) or macrophages (anti-Mac-3) and the number of immunostained cells per high power field were quantified (A and B). Mean ± SE; * denotes P < 0.05 versus euthermic mice at the same time point. Because hyperthermic mice did not survive beyond 60 hours of exposure to hyperthermia, analysis of later time points could not be done (ND). Panels C and D show representative PMN-immunostained lung sections from hyperthermic (C) and euthermic (D) mice exposed to F1O2 > 0.95 for 36 hours. Arrows indicate PMN within the alveolar septal wall. Magnification, ×400.

Figure 6.

Effect of febrile range hyperthermia on inflammatory cell accumulation in BALF. Euthermic and hyperthermic mice were exposed to FIO2 > 0.95. Groups of eight mice were sacrificed at the indicated time, lungs were lavaged, and the numbers of PMN (A) and macrophages (B) in BALF determined by manual counting. Mean ± SE; * denotes P < 0.05 versus euthermic normoxic controls (time 0). ** denotes P < 0.05 versus euthermic mice at the same exposure time. Because hyperthermic mice did not survive beyond 60 hours of exposure to hyperthermia, analysis of later time points could not be done (ND).

Figure 7.

Effect of FRH on numbers of circulating leukocytes and plasma concentration of G-CSF. Groups of five mice were exposed to FIO2 0.21 (normoxic) or > 0.95 (hyperoxic) and housed at either 24°C (euthermic) or 34.5°C (hyperthermic). Mice were sacrificed after 24 hours of exposure and heparinized blood was collected via cardiac puncture and analyzed for number of leukocytes (A) and PMN (B) by manually counting. C: Plasma concentration of G-CSF was measured by ELISA. Mean ± SE; * denotes P < 0.05 versus euthermic, normoxic mice; † denotes P < 0.05 versus euthermic, hyperoxic mice. D: Effect of G-CSF immunoblockade on circulating PMN count in mice exposed to hyperthermia for 24 hours. Mice were pretreated with anti-G-CSF or control rabbit serum (sham) 2 hours before beginning 24 hours of euthermic or hyperthermic exposure; ** denotes P < 0.05 versus sham-treated/euthermic; ‡ denotes P < 0.05 versus sham-treated/hyperthermic mice.

Figure 8.

Effect of febrile range hyperthermia on ELR+ CXC expression in the hyperoxic lung. A: Groups of eight mice were exposed to FIO2 0.21 or > 0.95 and housed at either 24°C (euthermic) or 34.5°C (hyperthermic) for the indicated time. Lung lavage was performed and KC was quantified by ELISA. * denotes P < 0.05 versus euthermic, normoxic controls. ** denotes P < 0.05 versus euthermic, hyperoxic mice at the same exposure time. Because hyperthermic mice did not survive beyond 60 hours of exposure to hyperthermia, analysis of later time points was not done (ND). B and C: BALF from groups of five mice exposed to FIO2 0.21 or > 0.95 and housed at either 24°C (euthermic) or 34.5°C (hyperthermic) for 24 hours were analyzed for concentrations of LIX (B) and MIP-2 (C) using ELISA. Mean ± SE; * denotes P < 0.05 versus euthermic normoxic mice.

Figure 9.

Effect of treatment with anti-CXCR2 antiserum. Mice received goat anti-mouse CXCR2 antiserum, preimmune goat serum (control), or no pretreatment (untreated) 2 hours before transfer into a 34.5°C chamber and exposure to FIO2 > 0.95 and sacrificed 40 hours later. A: The lungs were lavaged and the number of PMN in BALF was determined in groups of 10 mice. B: Lungs were fixed at 20 cm H2O, immunostained with anti-GR-1 antibody, and the number of PMN per HPF were counted in groups of four mice. Untreated, euthermic, air-breathing (FIO2 > 0.95) mice were analyzed as a comparison (hatched bar). * denotes P < 0.05 versus untreated and pre-immune serum-treated control mice. C: Representative micrographs from sham-treated and anti-CXCR2-treated mice. Arrows indicate PMN. D: Median survival was determined in groups of nine mice.