Published ahead of print on April 1, 2004, doi:10.1164/rccm.200306-846OC
© 2004 American Thoracic Society
Pulmonary Neutrophil Infiltration in Murine SepsisRole of Inducible Nitric Oxide SynthaseVascular Biology Group, Lawson Health Research Institute, Division of Respirology, Departments of Medicine, Physiology, and Pharmacology, London Health Sciences Center, University of Western Ontario, London, Ontario, Canada Correspondence and requests for reprints should be addressed to Sanjay Mehta, M.D., F.R.C.P.C., F.C.C.P., Division of Respirology, London Health Sciences Center South Street Campus, 375 South Street, London, ON, N6A 4G5 Canada. E-mail: sanjay.mehta{at}lhsc.on.ca
Nitric oxide (NO) derived from inducible NO synthase (iNOS) contributes to the pathophysiology of acute lung injury (ALI). The effect of iNOS on pulmonary neutrophil infiltration in ALI is not known. Thus, we assessed pulmonary microvascular neutrophil sequestration through intravital videomicroscopy and pulmonary neutrophil infiltration, reflected by myeloperoxidase activity and lavage neutrophil counts, after induction of sepsis by cecal ligation/perforation in wild-type (iNOS+/+) versus iNOS–/– mice. Pulmonary microvascular neutrophil sequestration was attenuated in septic iNOS–/– versus iNOS+/+ mice (15 ± 1 vs. 20 ± 1 leukocytes per field, p < 0.05), but lavage neutrophil counts were greater in iNOS–/– mice (5.7 ± 1.5% vs. 0.7 ± 0.1%, p < 0.05) between 6 and 18 hours after cecal ligation and perforation. When iNOS+/+ bone marrow was transplanted into bone marrow–depleted iNOS–/– mice (+ to – chimeras; iNOS limited to marrow-derived inflammatory cells), septic pulmonary microvascular neutrophil sequestration and lavage neutrophil counts were restored to levels seen in septic iNOS+/+ mice. In contrast, in – to + chimeras, pulmonary neutrophil trafficking was similar to iNOS–/– mice. In vitro cytokine-stimulated neutrophil transendothelial migration was significantly greater for iNOS–/– versus iNOS+/+ neutrophils (7.9 ± 0.7% vs. 3.8 ± 0.6%, p < 0.05) but was independent of endothelial iNOS. Thus, neutrophil iNOS-derived NO is an important autocrine modulator of pulmonary neutrophil infiltration in murine sepsis.
Key Words: sepsis • acute lung injury • neutrophil infiltration • inducible nitric oxide synthase • reciprocal bone marrow transplant chimeras Sepsis-induced acute lung injury (ALI) remains a major clinical problem with significant morbidity and mortality (1–3). Polymorphonuclear leukocytes (neutrophils) are thought to contribute significantly to the pathophysiologic features of ALI, such as the exudation of protein-rich fluid across the alveolocapillary endothelial cell barrier (4–8). A pathologic hallmark of ALI is the pulmonary microvascular sequestration and subsequent tissue infiltration of neutrophils (9, 10). Enhanced pulmonary neutrophil sequestration and infiltration in sepsis and ALI are the net result of changes in key neutrophil functions. These changes include reduced neutrophil deformability, increased neutrophil surface expression and activation of cell–cell adhesion molecules, and enhanced release of soluble mediators, including proteolytic enzymes and reactive oxygen species (11–13). ALI is also characterized by upregulation of inducible nitric oxide (NO) synthase (iNOS) and increased production of NO in both animal models and in humans (14–16). This iNOS-derived NO contributes to the pathophysiologic features of ALI (8, 17–19). Moreover, iNOS-derived NO may also have complex effects on alveolar fluid clearance in ALI, inhibiting clearance early but increasing clearance later after intratracheal endotoxin (20). The majority of pulmonary cell types, including parenchymal cells (e.g., epithelial, endothelial) and bone marrow (BM)-derived inflammatory cells (e.g., neutrophils, macrophages), can release iNOS-derived NO (21, 22). However, there are few data on differential effects of iNOS-derived NO in ALI from these different cellular sources. Our recent work using reciprocal BM transplantation (BMT) to generate iNOS chimeric mice has defined cell population-specific quantitative differences in iNOS activity and NO production in endotoxin-induced ALI (23). Moreover, we demonstrated that protein-rich pulmonary edema and pulmonary oxidant and nitrosative stress in mice with septic ALI were absolutely dependent on inflammatory cell iNOS with no obvious contribution from iNOS in pulmonary parenchymal cells (8). As well, an immunohistochemical study of inflammatory bowel disease has suggested that cell-source–specific nitrosative and oxidative effects of iNOS-derived NO may also be important in humans (24). NO may also directly affect neutrophil function in ALI. For example, iNOS-derived NO decreased endotoxemic systemic and pulmonary leukocyte infiltration (25). NO can attenuate the upregulation of adhesion molecules in stimulated neutrophils (26). As well, others and we have shown that inhaled NO can reduce pulmonary neutrophil sequestration and infiltration in various models of ALI (14, 27). However, the effect of iNOS-derived NO in general and specifically neutrophil iNOS on pulmonary neutrophil sequestration and infiltration in ALI is not known. Thus, we investigated the role of iNOS-derived NO in sepsis-induced pulmonary neutrophil infiltration. Based on our previous observations of an important role of neutrophil iNOS in the pathophysiologic features of septic ALI, we hypothesized that neutrophil iNOS may facilitate septic pulmonary neutrophil infiltration. To test this hypothesis, we assessed pulmonary microvascular neutrophil sequestration using pulmonary intravital videomicroscopy (IVM), as well as pulmonary tissue and alveolar neutrophil infiltration in septic wild-type (iNOS+/+) versus iNOS–/– mice. Furthermore, reciprocal BMT iNOS chimeric mice were used to characterize the differential contribution of iNOS in discrete cell populations to modulation of septic neutrophil infiltration. Finally, we isolated neutrophils and pulmonary microvascular endothelial cells (PMVECs) from iNOS+/+ and iNOS–/– mice to assess directly the action of neutrophil iNOS on transendothelial neutrophil migration in vitro. Some of these data have previously been reported in abstract form (28, 29).
The institutional animal research committee approved all studies. Male iNOS+/+, iNOS–/– C57Bl/6 mice (8–10 weeks, 25–30 g; Charles River, St. Constant, PQ, Canada), and BMT mice were randomized to sham versus volume-resuscitated cecal ligation and perforation (CLP) sepsis under halothane anesthesia (8, 14, 30). At various time points, mice were prepared for pulmonary IVM or were sacrificed (pentobarbital) for bronchoalveolar lavage (BAL) or collection of blood and lung tissue. BMT iNOS chimeras were generated as previously described (8, 23). Irradiated, BM-depleted mice were BM reconstituted and studied 4 weeks later. Two chimeras were generated: + to – (iNOS+/+ BM cells transplanted into iNOS–/– recipient) and – to +. A control + to + BMT mouse was also generated. Details are provided in the online supplement.
Blood–Pulmonary Neutrophil Trafficking
Pulmonary IVM
In Vitro Neutrophil Migration
105 PMVECs were grown to confluence on 0.1% gelatin-coated cell culture inserts (3-µm diameter pores). Five x 105 neutrophils were applied to the apical aspect of the PMVEC monolayer for 3 hours in the presence of 30-pg/mL IFN-
Statistical Analysis
The Effect of iNOS on Sepsis-induced Pulmonary Neutrophil Sequestration and Infiltration After CLP, septic mice exhibited lethargy, decreased exploratory behavior, and tachypnea. In this volume-resuscitated septic model (30), we have previously reported stable mean arterial pressure over 18–24 hours (14). We have also previously shown that the severity of peritonitis post-CLP is similar in iNOS+/+ versus iNOS–/– mice, as reflected by peritoneal lavage bacterial counts, cellularity, and protein concentration (8). There were no differences in total white blood cell (2.6 ± 0.2 vs. 2.1 ± 0.6 x 109/L, respectively, p = NS) or absolute neutrophil count (0.33 ± 0.03 vs. 0.32 ± 0.07 x 109/L, respectively, p = NS) between naive iNOS+/+ and iNOS–/– mice. Sepsis was associated with a similar increase in blood neutrophils (as a percentage of total white blood cell) in both iNOS+/+ and iNOS–/– mice at 2 hours after CLP (Figure 1) . At this time, the absolute blood neutrophil count was also significantly increased in both septic iNOS+/+ and iNOS–/– mice (1.19 ± 0.34 vs. 1.16 ± 0.25 x 109/L, respectively, p < 0.05 for both vs. respective naive), whereas the total white blood cell count was largely unchanged (2.2 ± 0.4 vs. 2.0 ± 0.3 x 109/L, respectively). This sepsis-induced increase in blood neutrophils remained significant in iNOS+/+ mice at 6 and 18 hours after CLP, but was no longer different from baseline in iNOS–/– mice at 18 hours. There were no differences in total white blood cell counts between septic iNOS+/+ and iNOS–/– mice at any time point (data not shown). Blood hematocrit was similar in septic iNOS+/+ and iNOS–/– mice (0.44 ± 0.02 vs. 0.45 ± 0.02, respectively, n = 4 for each, p = NS) at 6 hours after CLP. Moreover, hematocrit values in these septic mice were not different from naive mice (data not shown) or from published values in wild-type C57Bl/6 mice (33).
As assessed by pulmonary IVM, sepsis was associated with significant pulmonary microvascular leukocyte sequestration in iNOS+/+ mice (Figure 2) . Video recordings in naive and septic iNOS+/+ mice are provided in the online supplement (Figures E1 and E2). Assessment of the time course of pulmonary microvascular leukocyte sequestration by IVM revealed significant sepsis-induced sequestration in both iNOS+/+ and iNOS–/– mice at all time points from 1 to 18 hours after CLP (Figure 3) . However, there were significantly fewer sequestered leukocytes in iNOS–/– versus iNOS+/+ mice at all time points (p < 0.05).
Sepsis was also associated with a significant increase in pulmonary MPO activity in both iNOS+/+ and iNOS–/– mice at 1 hour after CLP (Figure 4) . This sepsis-induced increase in MPO persisted throughout the 18-hour time course studied, with no difference between iNOS+/+ and iNOS–/– groups.
Neutrophil counts in BAL were significantly increased in both iNOS+/+ and iNOS–/– septic mice 1 hour after CLP (Figure 5) . However, from 2 to 18 hours after CLP, BAL neutrophil counts remained elevated in only iNOS–/– mice, whereas there were no differences between septic and naive iNOS+/+ mice. BAL return volumes were similar in iNOS+/+ and iNOS–/– mice. Furthermore, the total BAL cell count was not different between septic iNOS+/+ and iNOS–/– mice (data not shown).
The Effect of Selective Presence of iNOS in Inflammatory Versus Parenchymal Cells on Sepsis-induced Pulmonary Neutrophil Sequestration and Infiltration As we have previously reported (8, 23), BMT iNOS chimeric mice permit the differential effects of different cellular sources of iNOS in sepsis-induced pulmonary neutrophil infiltration to be determined. iNOS+/+ recipient mice transplanted with iNOS+/+ BM (+ to + group) are genotypically identical to wild-type iNOS+/+ mice but serve as a control for all transplantation procedures. In naive + to + mice, blood neutrophil counts, pulmonary leukocyte sequestration by IVM, and BAL neutrophil counts were similar to values in naive wild-type iNOS+/+ mice (data not shown). At 18 hours after CLP, sepsis-induced increases in the percentage of blood neutrophils, pulmonary microvascular leukocyte sequestration, and pulmonary MPO activity (data not shown) were similar in + to + BM-transplanted mice (Figure 6) , as described previously here for nontransplanted iNOS+/+ mice. In addition, as in iNOS+/+ mice, there was no increase in the percentage of BAL neutrophils at 18 hours after CLP in septic + to + mice versus naive + to + mice. Thus, the BMT protocol (lethal irradiation-induced BM depletion and subsequent BM reconstitution) did not significantly affect sepsis-induced pulmonary neutrophil sequestration and infiltration.
In naive (nonseptic) mice, blood neutrophil counts, pulmonary microvascular sequestered leukocytes by IVM, and BAL neutrophil counts were similar between + to – and – to + iNOS chimeric groups and were not different from naive + to + mice. Thus, data were combined across the three naive BMT groups and are reported as single values for each of these three parameters (n = 15; Figure 6). In + to – iNOS chimeras (iNOS limited to donor BM-derived inflammatory cells), sepsis-induced changes in blood neutrophil counts, pulmonary microvascular leukocytes sequestration, and the percentage of BAL neutrophils 18 hours after CLP were similar to septic + to + mice (Figure 6). In sharp contrast to septic + to + and + to – mice, CLP in – to + iNOS chimeric mice (iNOS limited to recipient tissue stromal cells) resulted in no increase in blood neutrophils, a lesser degree of pulmonary leukocyte sequestration by IVM, but markedly greater percentage of BAL neutrophils. Furthermore, these responses in pulmonary microvascular leukocyte sequestration and blood and BAL neutrophil counts in septic – to + iNOS chimeric mice were similar to those described previously here in iNOS–/– mice.
Effect of Neutrophil Versus Endothelial Cell iNOS on In Vitro Transendothelial Neutrophil Migration
The migration of unstimulated iNOS+/+ and iNOS–/– neutrophils across gelatin-coated transwell inserts, in the absence of PMVECs, was identical (11.7 ± 1.8% vs. 14.0 ± 3.0%, p = NS). Similarly, there was no significant difference in migration of iNOS+/+ and iNOS–/– neutrophils in the presence of cytomix, but in the absence of PMVECs (12.3 ± 2.2% vs. 13.5 ± 2.3%, p = NS).
In the volume-resuscitated murine CLP model of sepsis used in this study, ALI is characterized by increased MPO activity, a marker of neutrophil infiltration, oxidant stress, high-protein pulmonary edema, and increased expression and activity of iNOS (8, 14). In this study, we further explored the kinetics of sepsis-induced changes in blood neutrophil counts, pulmonary microvascular neutrophil sequestration, as well as pulmonary tissue and alveolar neutrophil infiltration. Sepsis was associated with increased circulating blood neutrophils, pulmonary microvascular neutrophil sequestration, increased pulmonary MPO activity, and increased BAL neutrophil counts in both iNOS+/+ and iNOS–/– mice. However, key differences in septic iNOS–/– mice include a greater decline in circulating blood neutrophils at 18 hours after CLP, lesser pulmonary microvascular neutrophil sequestration, and increased BAL neutrophil counts versus iNOS+/+ mice. Moreover, sepsis-induced changes in blood and pulmonary neutrophil counts in iNOS–/– mice were reproduced in – to + BM-transplanted iNOS chimeric mice, in which iNOS is still present in recipient stromal cells but absent from donor BM-derived inflammatory cells. In contrast, septic blood and pulmonary neutrophil responses in + to – iNOS chimeras (iNOS only in donor BM-derived inflammatory cells) were similar to iNOS+/+ mice. Thus, absence of iNOS specifically in BM-derived inflammatory cells (e.g., neutrophils and macrophages) was associated with lower blood and pulmonary microvascular neutrophil counts in the presence of increased alveolar neutrophil counts in septic mice. Finally, in vitro studies indicated that the specific absence of iNOS in neutrophils was associated with increased cytokine-stimulated transendothelial neutrophil migration, independent of endothelial iNOS presence/absence. Moreover, the enhanced migration of iNOS–/– neutrophils was dependent on the presence of an endothelial barrier, suggesting that an interaction between neutrophil iNOS-derived NO and endothelial cells retards transendothelial neutrophil migration.
Effect of NO on Pulmonary Neutrophil Sequestration and Infiltration in ALI NO may influence pulmonary inflammatory responses through several mechanisms, including modulation of the microvascular sequestration and tissue infiltration of neutrophils (25, 26, 38). For instance, others and we have shown that the administration of exogenous, inhaled NO reduces pulmonary neutrophil infiltration in ALI due to various insults, including sepsis (14, 27). Similarly, NOS inhibition and reduced endogenous NO production are associated with enhanced rolling and adhesion of neutrophils in the systemic microcirculation (26, 39, 40). A role for endogenous NO in the modulation of pulmonary neutrophil infiltration has been controversial. In some studies of ALI, the absence of iNOS in iNOS–/– mice had no effect on neutrophil infiltration, as reflected by pulmonary tissue MPO activity (8, 41). In contrast, pulmonary MPO activity was significantly greater in lipopolysaccharide-treated iNOS–/– versus iNOS+/+ mice (25). In different models of ALI, non–isoform-selective inhibition of NOS by L-NAME either increased or decreased pulmonary neutrophil sequestration (42, 43). In this study, sepsis in iNOS–/– mice resulted in a significantly greater presence of neutrophils in the bronchoalveolar pulmonary compartment and reduced blood and pulmonary microvascular neutrophil counts compared with septic iNOS+/+ mice. Thus, iNOS deficiency in septic mice appears to facilitate the infiltration of neutrophils into pulmonary tissue from the pulmonary microvasculature. However, pulmonary MPO activity was increased similarly in both septic iNOS+/+ and iNOS–/– mice, concealing the significant differences in sequestered and infiltrated neutrophils. MPO in lung homogenate may be a global measure of pulmonary neutrophil presence, including microvascular sequestered neutrophils, as well as neutrophils that have infiltrated pulmonary interstitial and bronchoalveolar compartments. Pulmonary tissue MPO would not resolve differences in neutrophil presence in the various compartments. Alternatively, recent evidence suggests important endothelial cell endocytosis of free MPO released by the intravascular degranulation of neutrophils (44, 45). Thus, the pulmonary tissue MPO signal may partly reflect systemic neutrophil activation, making it less representative of tissue neutrophil infiltration.
Cell-source–dependent Effects of iNOS in the Pathophysiologic Features of Sepsis-induced ALI In this study, we used iNOS chimeras to define the cellular sources of iNOS responsible for the observed differences in blood, pulmonary microvascular, and bronchoalveolar neutrophil counts in septic iNOS+/+ versus iNOS–/– mice. Thus, as in iNOS–/– mice with complete absence of iNOS, the specific absence of iNOS in donor BM-derived inflammatory cells (e.g., neutrophils, macrophages) in – to + chimeras was sufficient to enhance sepsis-induced pulmonary neutrophil infiltration. Similarly, the specific presence of iNOS in inflammatory cells in + to – iNOS chimeras was sufficient to attenuate septic pulmonary neutrophil infiltration. This suggested a specific effect of iNOS in either neutrophils or macrophages on the transendothelial migration of neutrophils in septic ALI in vivo. Based on the results of the in vitro studies, performed in the absence of macrophages, we conclude that transendothelial neutrophil migration is attenuated by the specific presence of neutrophil iNOS. We cannot exclude an additional role for macrophage iNOS, as macrophages and neutrophils are both reconstituted after BMT, such that they share the same iNOS genotype in reciprocal iNOS chimeras (23). Indeed, an important interdependence between monocyte and neutrophil pulmonary trafficking was recently recognized as being essential for increased vascular permeability after intratracheal CC chemokine ligand 2 and endotoxin (46). It should be noted that transendothelial neutrophil migration in vitro was independent of endothelial cell iNOS genotype, consistent with our in vivo observations of a lack of role of stromal cell iNOS in septic pulmonary neutrophil infiltration. The mechanism of neutrophil iNOS attenuation of transendothelial neutrophil migration remains uncertain. However, the effect of neutrophil iNOS on neutrophil migration was not observed in the absence of endothelial cells and thus was endothelium dependent in vitro. As such, it is unlikely that the effect of iNOS was mediated through NO-dependent inhibition of neutrophil F-actin assembly and improved neutrophil deformability as previously reported (27). Moreover, NO-dependent improved neutrophil deformability might be expected to yield an increase in pulmonary microvascular neutrophil sequestration in iNOS–/– mice, rather than the observed decrease. It is possible that neutrophil iNOS-derived NO inhibited expression of adhesion molecules on either the neutrophil (e.g., CD18) or endothelial cell (e.g., intercellular adhesion molecule-1) surface, modulated neutrophil oxidant or protease release, or influenced interendothelial cell gap formation (26, 47–49). Further in vivo and in vitro studies will address these possibilities. The clinical relevance of murine studies is moot. Although a mouse model can never fully recapitulate the complex human condition of ALI, a disease model, such as CLP, is most analogous to clinical human disease and is clearly better than models employing injection of lipopolysaccharide. In addition, our murine model of sepsis is associated with mild ALI, as reflected by the minimal BAL neutrophilia, in stark contrast to the marked pulmonary neutrophil influx in human ALI. The advantage of such a mild model is the lack of confounding treatments, such as mechanical ventilation and high levels of supplemental oxygen, which may clearly contribute to ALI (50). As such, this model permits a focus on the earliest, initiating events in sepsis-induced ALI, which are likely consistent between species. These early changes occur before the development of severe ALI or other confounding features, such as secondary infection or activation of reparative mechanisms. We believe it is at these earliest time points of the natural history of ALI that patients will optimally benefit from novel therapeutic strategies to target inflammatory events. Finally, although it is clear that iNOS contributes significantly to ALI in rodents, the role of iNOS and NO in human ALI remains an important, unsettled question. Increased iNOS expression and/or activity have been reported in inflammatory cells isolated from septic humans (15, 51, 52). Moreover, increased immunohistochemical staining for 3-nitrotyrosine, a marker of NO-dependent oxidative stress, is found in cells and tissues from humans with sepsis and ALI (53–56). Clearly, further research specifically with human material from patients with ALI, as well as in vitro studies of human PMVEC–neutrophil interaction will shed more light on the role of iNOS in human ALI. In summary, in a murine model of sepsis-induced ALI, neutrophil iNOS appears to retard the transendothelial migration and pulmonary infiltration of neutrophils. This effect of neutrophil iNOS was endothelium dependent, suggesting an interaction between neutrophil iNOS-derived reactive nitrogen species and endothelial cells in modulating transendothelial neutrophil migration. Moreover, we do recognize a paradox in our current and recent observations. iNOS–/– neutrophils infiltrate the lung more readily under septic conditions but do not induce pathophysiologic features of sepsis-induced lung injury, that is, high-protein edema and pulmonary oxidant stress. We speculate that the retarded transendothelial migration of iNOS+/+ neutrophils enhances local release of NO and reactive nitrogen species at the microvascular endothelial barrier, enhancing injury. Thus, measures to enhance pulmonary neutrophil infiltration may paradoxically be therapeutically beneficial, possibly via accelerated neutrophil apoptosis.
Supported by the Canadian Institute of Health Research, Ontario Thoracic Society. This article has an online supplement, which is accessible from this issue's table of contents online at www.atsjournals.org Conflict of Interest Statement: H.M.R. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; L F.W. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; S.W. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; M.R. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; C.L. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; D.G.M. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; S.M. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. Received in original form June 24, 2003; accepted in final form March 31, 2004
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
METHODS
, tumor necrosis factor-
, and interleukin-1ß (Sigma) in both upper and lower compartments. Trans-PMVEC neutrophil migration (percentage of neutrophils applied to the apical surface) was quantified by counting stained neutrophils in the basal compartment. 
