NOX2 protects against prolonged inflammation, lung injury, and mortality following systemic insults

Abstract

The systemic inflammatory response syndrome (SIRS) is a clinical condition occurring in intensive care unit patients as a consequence of both infectious and noninfectious insults. The mechanisms underlying resolution of SIRS are not well characterized. NOX2 (NADPH oxidase 2)-derived reactive oxygen species are critical for killing of certain pathogens by polymorphonuclear leukocytes (PMN). Patients with chronic granulomatous disease who lack functional NOX2 are not only prone to serious infections, they also exhibit chronic inflammatory conditions, suggesting a local anti-inflammatory role for NOX2. We hypothesized that NOX2 is required for the resolution of sterile systemic inflammation. Using a murine model of sterile generalized inflammation, we observed dramatically increased mortality of gp91(phox-/y) (NOX2-deficient) as compared to wild-type (WT) mice. Both genotypes developed robust SIRS with hypothermia, hypotension, and leukopenia; however, WT mice recovered within 48 h whereas NOX2-deficient mice did not. Although both groups displayed rapid peritoneal PMN recruitment, the recruited NOX2-deficient PMN demonstrated an enhanced inflammatory phenotype. Moreover, NOX2-deficient mice exhibited a hemorrhagic inflammatory response in the lungs with rapid and persistent recruitment of neutrophils to the alveolar space, whereas WT mice had minimal lung pathology. Several proinflammatory cytokines remained elevated in NOX2-deficient mice. The persistent inflammatory environment observed in NOX2-deficient mice resulted from continued peritoneal chemokine secretion and not delayed apoptosis of PMN. These data suggest a requirement for NOX2 in the resolution of systemic inflammation.

Fig. 1

NOX2-deficient mice exhibit increased mortality following zymosan injection. Kaplan-Meier curve displaying survival of WT and NOX2-deficient mice following zymosan injection. Injection of 0.7 mg/g i.p. zymosan elicited markedly increased mortality in NOX2-deficient mice (n = 11 per genotype; ** p ≤ 0.01), whereas a 30% reduction in the zymosan dose elicited minimal mortality in either genotype (n = 1/9 WT, 0/8 gp91^phox-/y).

Fig. 2

NOX2-deficient mice develop prolonged hypothermia and hypotension following zymosan injection. Core body temperature (a) and BP (b) were measured at several time points following induction of SIRS by zymosan injection. a Mean core body temperature ± SEM (n ≥ 7 animals per genotype per time point from 2 independent experiments). Two-way ANOVA with Bonferroni post tests revealed statistical significance between genotypes 6, 24, 30, and 48 h after injection. *** p < 0.001. b Mean arterial pressure ± SEM (n ≥ 10 per genotype per time point from 3 independent experiments). Data were analyzed by two-way ANOVA with Bonferroni post tests. * p < 0.05; ** p < 0.01.

Fig. 3

NOX2-deficient mice have prolonged leukopenia and thrombocytopenia following induction of SIRS. Complete blood counts were performed on noninjected mice (time 0), and 6, 24, and 48 h after zymosan injection to evaluate WBC concentration (a), neutrophil (b), lymphocyte (c), and monocyte percentages (d), hemoglobin concentration (e), and platelet count (f). Means + SEM (n = 4 mice per genotype at time 0, n ≥ 6 per genotype at subsequent time points from a minimum of 2 independent experiments). ** p < 0.01; *** p < 0.001.

Fig. 4

NOX2-deficient mice display more severe symptoms of systemic inflammation. A total 24-hour inflammation score (a) was generated by scoring saline- and zymosan-injected mice for symptoms of systemic inflammation including lethargy (b), ruffled fur (c), conjunctivitis (d), and diarrhea (e) 24 h after injection. Each dot represents a single mouse with means indicated by horizontal lines (n = 6 WT saline, n = 20 WT zymosan, n = 7 gp91^phox-/y saline, and n = 19 gp91^phox-/y zymosan from 3 independent experiments). * p < 0.05; ** p < 0.01; *** p < 0.001.

Fig. 5

Differential peritoneal PMN recruitment and PMN activation states in WT and NOX2-deficient mice. Flow cytometry was used to determine the percentage of PMN (Ly-6G+ cells) in the peritoneum and bone marrow of WT and NOX2-deficient mice as well as measure PMN surface expression of CXCR2, RAGE, and CD11b. a PMN recruitment to the peritoneum of noninjected mice (time 0) and zymosan-injected mice 6, 24, and 48 h after injection. Means + SEM (n = 4 noninjected and n ≥ 9 zymosan-injected mice per genotype per time points from a minimum of 3 independent experiments). *** p < 0.0001. NS = Nonsignificant. b Percentage of PMN in the bone marrow of noninjected mice (time 0) and zymosan-injected mice 6, 24, and 48 h after injection. Means + SEM (n = 4 noninjected and n ≥ 13 zymosan-injected mice per genotype from a minimum of 3 independent experiments). * p < 0.05. c-j The geometric mean index (GMI) of several inflammatory surface markers was determined by flow cytometry for gated PMN (Ly-6G+) populations isolated from the peritoneum and bone marrow of zymosan-injected mice 6, 24, and 48 h after injection. GMI of peritoneal PMN for CXCR2 (c), RAGE (f), and CD11b (i). GMI of bone marrow PMN for CXCR2 (d), RAGE (g), and CD11b (j). Fold increase (peritoneal PMN GMI/bone marrow PMN GMI) of CXCR2 (e), RAGE (h), and CD11b (k). Means + SEM (n ≥ 9 per genotype per time point from a minimum of 3 independent experiments). * p < 0.05; ** p < 0.01; *** p < 0.001.

Fig. 6

NOX2-deficient mice have enhanced lung pathology compared to WT mice. a Percentage of PMN in the cells recovered from the BAL fluid of mice sacrificed 6, 24, and 48 h after zymosan injection. Hemorrhage (b), neutrophil infiltration (c), and thrombi (d) were scored in a blinded fashion 6, 24, and 48 h after zymosan injection (24 h data shown), representative images were taken (e-g), and total lung pathology scores were generated (h). a-d Means + SEM. * p < 0.05; ** p < 0.01; *** p < 0.001. a n ≥ 7 per genotype per time point from 3 independent experiments. b-d n = 2 saline- and 11 zymosan-injected mice per genotype from 3 experiments. e-g Representative images (magnification: ×600) of alveolar sections taken 24 h after zymosan injection. e Immune cell infiltration in a NOX2-deficient lung. f Hemorrhage and a thrombus in a NOX2-deficient lung. g WT lung (control). h Means + SEM (n = 6-11 per genotype per time point from a minimum of 3 independent experiments). ** p < 0.01; **** p < 0.0001.

Fig. 7

Plasma levels of MCP-1 (a), G-CSF (b), RANTES (c), IL-17 (d), MIP-1β (e), IL-6 (f) and IL-10 (g) were persistently elevated in NOX2-deficient mice. Time 0 = Noninjected. For MCP-1, G-CSF, and IL-6 some values were out of range for the multiplex ELISA. High out of range values were assigned the peak value within the measurable range for statistical comparison. Means + SEM (n ≥ 7 noninjected and n ≥ 9 zymosan-injected mice per genotype per time point from 3 independent experiments). * p < 0.05; ** p < 0.01.

Fig. 8

Increased frequency of IL-17A-producing T-cell subsets in NOX2-deficient mice. Flow cytometry was used to identify CD4+, IL-17A+, and γδ TCR+ cells in the peritoneum of zymosan-injected mice 24 h after injection. a Percentage of CD4+ cells. b CD4 T cells were gated and the percentage that was IL-17A+ was determined. c Percentage of γδ TCR+ cells. d γδ T cells were gated and the percentage that was IL-17A+ was determined. Means + SEM (n = 7 per genotype from 2 independent experiments). * p < 0.05; *** p < 0.001.

Fig. 9

Peritoneal PMN from NOX2-deficient mice display increased apoptosis. Flow cytometry was used to determine the percentage of PMN undergoing apoptosis in the peritoneum 24 and 48 h after SIRS induction. Apoptosis was evaluated by annexin V staining for exposed phosphatidylserine and Hoechst staining. Percentage of live (a; annexin V-negative, Hoechst-negative) and apoptotic (b; annexin V-positive) peritoneal PMN. Means + SEM (n ≥ 10 per genotype per time point). * p < 0.05; ** p < 0.01; *** p < 0.001.

Fig. 10

Persistently elevated levels of inflammatory mediators in the peritoneum of NOX2-deficient mice. IL-6 (a), KC (b), MCP-1 (c), MIP-1α (d), and MIP-1β (e) concentrations were measured in peritoneal fluid recovered from WT and NOX2-deficient mice 6, 24, and 48 h after SIRS induction. Means + SEM (n ≥ 6 per genotype per time point from 3 independent experiments). * p < 0.05; ** p < 0.01.