This study tested the hypothesis that the unique phenotype of alveolar macrophages (AM) is maintained through adaptation to the relatively high oxygen partial pressure (Po2) of the lung, through modification of redox-sensitive transcription factors. BALB/c mouse bone marrowderived macrophages (BMC) were differentiated under different Po2 and compared functionally to AM and peritoneal macrophages (PM). BMC differentiated in normoxia (Po2 140 Torr, BMChigh) were similar to AM in having low phagocytic and antigen presenting cell (APC) activities. However, BMC grown in low oxygen tension as found in other tissues (<40 Torr, BMClow) were better phagocytes and APCs, similar to PM. BMChigh were more oxidative intracellularly than BMClow, based on oxidation of dichlorofluorescein and higher glutathione disulfide/glutathione (GSH) ratios, despite having more GSH. Finally, lipopolysaccharide-induced nuclear factor-κB translocation, measured by laser scanning cytometry, was reduced in BMChigh and AM, compared with BMClow and PM, respectively. These data suggest that regulation of the AM phenotype may occur, at least in part, via inhibition of NF-κB by the unique redox environment.
- alveolar macrophage
- nuclear factor-κB
due to its large surface area exposed to the environment, the lung contends with unique challenges on a constant basis and is protected from damage by respired microbes and other particles through unique immunomodulatory capabilities. However, the prevalence and severity of chronic lung inflammatory disorders, such as asthma and fibrosis, demonstrate the susceptibility of this organ to dysfunction from which it often cannot recover. The lung maintains a fairly immunosuppressed environment, while retaining the ability to mount innate and specific immune responses. This allows most contaminants to be cleared by mechanical means, avoiding constant inflammatory reactions that would result in chronic damage. The alveolar macrophage (AM) is in close contact with the environment, making it one of the major cells involved in both mechanical clearance and maintaining immune homeostasis. Compared with other tissue macrophages such as peritoneal macrophages (PM), AM have been shown to be relatively poor in antigen presenting activity (5, 14) and to actively suppress some dendritic cell (DC) and T cell activities (3, 10, 23). Studies have suggested that the factors mediating this suppression may include macrophage-derived products (2, 15, 19, 24) and low expression of the B7 costimulatory molecules for antigen presenting cell (APC) activity (3). However, there is very little information regarding how the AM is maintained in this immune regulatory state. Therefore, a study of AM phenotypic regulation is critical to an understanding of lung immunotoxicology.
Although many factors contribute to tissue-specific differentiation, one of the critical factors unique to AM might be the high oxygen tension present in the lung. Compared with other tissue macrophages, AM are exposed to much higher oxygen partial pressures (Po2), which might require compensatory mechanisms not needed in other tissues, such as increased production of antioxidants. In addition, the AM are phagocytic, leading to oxidative activities within the cells themselves. Because oxidative stress can be described as the inability of a cell's antioxidant systems to deal adequately with levels of oxygen and its radicals, prevention of oxidative stress in this environment requires a different antioxidant status from that in other tissues. In macrophages, the same transcription factors that respond to oxidative stress (for example NF-κB and activator protein-1) are also involved in the regulation of many macrophage functions, including production of inflammatory mediators (20, 21). Nevertheless, normal AM often seem to have an anti-inflammatory role. Therefore, a mechanism clearly exists to protect the cells from oxidative stress in the absence of true stress signaling, and this suggests that there are unique settings for the transcription factors involved in these activities.
As stated above, a major transcription factor in immune cells is NF-κB, which has been shown to be redox sensitive due to reactive thiol groups on signal pathway components (11). Genes that contain NF-κB response elements appear generally to be involved in proinflammatory activities or responses to stress. Several characteristics of bone marrow-derived DC that lead to highly immune active APC have recently been shown to be coordinately regulated through NF-κB (27). These include upregulation of major histocompatibility complex (MHC) II, B7 costimulatory molecules (CD80, CD86), and cytokines. This is consistent with the prevailing notion that NF-κB activity plays a critical role in macrophage function and suggests that this transcription factor might be part of the mechanism whereby redox changes could alter phenotype.
The hypothesis for this study was based on the proposal that AM must compensate for their toxic environment by upregulation of antioxidants to prevent constant oxidant stress, but this must occur without the normally coincident expression of proinflammatory mediators. By suppressing the latter, redox control of gene expression in the lung may be a critical factor in maintaining the AM phenotype. For this study, differentiation of murine bone marrow-derived macrophages (BMC) in different oxygen tension environments was used to demonstrate redox regulation of macrophage phenotypes, and that adaptation to extracellular oxygen tension is linked to transcriptional mechanisms that would affect phenotype. Ultimately, this new understanding of redox-dependent macrophage responsiveness will have implications for intervention in the pathological progression of lung inflammatory conditions.
Animals. BALB/c mice were obtained from Jackson Laboratories (Bar Harbor, ME), and DO11.10 [BALB/c background with transgene for ovalbumin (OVA)-specific T cell receptor (TCR)] breeding pairs were kindly provided by Corixa (Hamilton, MT). Euthanasia was performed by intraperitoneal injection of a lethal dose of pentobarbital sodium (for AM) or CO2 asphyxiation, which are both consistent with the recommendations of the Panel on Euthanasia of the American Veterinary Medical Association. The animal room was set on 12-h light/dark cycles at ∼18–26°C, with mouse feed and deionized water provided ad libitum. All protocols for the use of animals in experiments have been approved by the University of Montana Institutional Animal Care and Use Committee. The mice were maintained in microisolation containers in the Laboratory Animal Resources animal facility, in accordance with the “Guide for the Care and Use of Laboratory Animals” prepared by the Institute of Laboratory Animal Resources, National Research Council.
Collection and culture of BMC. The femurs of euthanized BALB/c mice were exposed, excised, and placed into a dish containing sterile PBS. In a sterile hood, the bone marrow cells were aspirated via a 1-ml syringe filled with culture media (RPMI 1640; Mediatech, Herndon, VA; 20% FBS; GIBCO, Grand Island, NY; 1% penicillin/streptomycin, GIBCO). Aspirated material was centrifuged and resuspended in fresh medium, counted, and seeded to tissue culture flasks or plates for experiments. After stromal cell elimination by adherence overnight, nonadherent cells were transferred to new flasks, and macrophage colony stimulating factor (M-CSF; R&D Systems, Minneapolis, MN) was added to give 20 ng/ml. The medium, with CSF, was replaced after 3–4 days. By 7–10 days the cells were fully differentiated and stained positive for F4-80 macrophage marker, as well as an antiaminopeptidase antibody (clone ER-BMDM1; Cedarlane, Hornby, Ontario, Canada) that is used for monitoring macrophage differentiation in culture. For challenge experiments, 10 μg/ml lipopolysaccharide (LPS, Salmonella typhimurium; Sigma, St. Louis, MO) was added for indicated times. Viability was determined to be >90% by trypan blue staining. Cell counts were performed on test cultures following differentiation and treatment with LPS to show that the cultures in the different incubators yielded similar numbers of cells.
Harvest of AM and PM. The lungs from euthanized BALB/c mice were surgically removed and thoroughly lavaged with five repeated 1-ml instillations of sterile PBS. We harvested peritoneal cells following CO2 asphyxiation by injecting 8 ml of sterile PBS into the peritoneal cavity following death and withdrawing the fluid into a sterile 10-ml syringe via an 18-gauge needle. The cells were kept on ice until centrifuged, resuspended, and counted on a Coulter counter (Z1 particle counter; Beckman Coulter, Miami, FL). To purify AM and PM by adherence, we plated the cells in RPMI 1640 with 10% FBS or mouse serum (preincubated in high or low oxygen for AM and PM, respectively) supplemented with antibiotics and placed them in the high- or low-oxygen 37°C cell culture incubator for ∼1 h. More than 95% of the resulting cells stained positive for the macrophage marker F4-80 (Caltag, Burlingame, CA) (data not shown).
Oxygen-controlled cell culture. To establish culturing conditions to mimic the oxygen tensions of alveolar spaces and other tissues, we set up two Thermo-Forma (Marietta, OH) tissue culture incubators side by side. One was left at the normal settings of 5% CO2 in ambient oxygen (21%), and the other was set to provide 5% CO2 and 5% O2 with N2 to flush out the excess oxygen. Using an Orion (Beverly, MA) model 830A dissolved oxygen meter, we have shown that in the 5% O2 incubator at 37°, RPMI culture medium containing 10% FBS developed a Po2 of ∼25 Torr, which corresponds with tissue values of <40 Torr (Fig. 1). This condition is referred to as “low oxygen” (BMClow) and approximates the Po2 that would be expected in many tissues of the body. In the regular 5% CO2 incubator, RPMI with FBS had a Po2 of 140–150 Torr, which falls between reported Po2 for atmospheric vs. alveolar conditions but is much higher than tissue Po2 (<40). Throughout this study, to distinguish this from the low-oxygen condition, we refer to this as the “high oxygen” condition (BMChigh), even though it is actually normoxic for the lung.
Dichlorofluorescein diacetate assay. Dichloro-dihydro-fluorescein diacetate (10 μM H2DCFDA; Molecular Probes, Eugene, OR) was added to the BMC in six-well culture plates for uptake, and after 1 h cells were washed with sterile PBS. The dye was oxidized gradually, with the rate dependent on the intracellular redox state. Cells without dye were used to subtract background. Readings for fluorescein fluorescence intensity were measured by flow cytometry (AM and PM) counting 104 cells in each sample, on a Becton-Dickinson FACSCalibur (San Jose, CA) or, for BMC, taken hourly for 3 h using a SpectraMax fluorescence plate reader set at 485-nm excitation/530-nm emission (Molecular Devices, Sunnyvale, CA).
Glutathione disulfide/glutathione analysis. Glutathione (GSH) levels and glutathione disulfide (GSSG)/GSH ratios were determined in a microtiter assay as described by Vandeputte et al. (25). Briefly, the cells were lysed in dilute HCl (10 mM), a small aliquot was removed for protein quantitation by Bradford protein assay from Pierce (Rockford, IL), and proteins were removed from the remaining lysate in 1.3% sulfosalicylic acid (SSA) precipitation. The supernatants were plated in 96-well plates and neutralized with a sodium phosphate/EDTA buffer, and then 5,5′-dithiobis-2-nitrobenzoic acid (DTNB) and NADPH were added at room temperature. The enzymatic reaction was started by addition of GSSG reductase. The plate was read immediately on a colorimetric plate reader for a kinetic analysis for 2 min at 405 nm with mixing. Final concentrations of the reagents were 0.73 mM DTNB, 0.24 mM NADPH, 0.09% SSA, and 1.2 IU/ml GSSG reductase. GSSG was measured after derivatization of GSH by 2-vinylpyridine before the above assay. Values were derived from standard curves.
Phagocytosis assay. To measure phagocytosis, we added fluorescent Escherichia coli particles (Phagocytosis Assay Kit, Molecular Probes) to the cultures that were plated at 5 × 105 cells/ml in eight-well chamber slides (Nalge-Nunc, Naperville, IL). The cells were allowed 1 h in their respective incubators for particle uptake, then were washed and analyzed by laser scanning cytometer (LSC; CompuCyte, Cambridge, MA) in which identical numbers of cells were counted (600 per well), and the mean/median integral fluorescence was determined.
APC assay. For APC activity, macrophages were cultured as above, treated with mitomycin C (Calbiochem, San Diego, CA) to prevent replication, washed, and then plated to 96-well tissue culture plates at 1 × 105 cells per well in 100 μl of RPMI with 20% FBS and antibiotics. OVA (Sigma) was added to give 8 mg/ml. T cells were obtained from DO11.10 mice that express an OVA-specific αβ TCR; their CD4+ T cells recognize an epitope of OVA presented on APC. Stimulation of DO11.10 T cells with antigen on APC induces proliferation and cytokine production without prior immunization. Spleens were removed from the mice, macerated between glass slides, and mixed well in a volume of PBS by pipetting to break up clumps. The cells were pelleted at 1,500 rpm for 5 min, and red blood cells were lysed in 10 ml of 0.83% ammonium chloride. The cells were then washed in PBS 3× to remove debris. T cells were enriched through a CD3 T cell enrichment column, according to the manufacturer's protocol (R&D Systems). The T cells were suspended in RPMI as above at 4 × 106 cells per ml and added to the wells containing macrophages in 100 μl per well to give 4 × 105 T cells/well. The plates were incubated for 24 h in their respective incubators, and then a bromodeoxyuridine (BrdU) label was added for the proliferation assay. After another 24 h, the plates were centrifuged (5 min at 1,500 rpm), and supernatants were removed for analysis of cytokine production by ELISA (Pharmingen, San Diego, CA). The protocol for fixing and staining the DNA was followed according to kit instructions (Oncogene, Boston, MA), and the plate was developed and read on a colorimetric plate reader at 450 nm.
NF-κB translocation by fluorescence microscopy/LSC. Cells to be analyzed were seeded to multiwell glass slides (Cel-Line) at equal density (5 × 104 cells/ml) and treated with media or media containing 10 μg/ml LPS for 1 h. The cells were fixed with 1% paraformaldehyde and then permeabilized with 0.2% Triton X-100. After blocking steps, the cells were stained with primary antibody, anti-NF-κB p65 subunit (Santa Cruz Biotechnology, Santa Cruz, CA), followed by AlexaFluor 488-labeled goat anti-rabbit IgG (Molecular Probes). The cells were counterstained with 5% propidium iodide (PI, Molecular Probes) with 100 μg/ml RNase to localize the nucleus. This staining protocol allowed visualization of the cellular location of NF-κB by confocal fluorescence microscopy with the oil-immersion ×60 objective, as well as quantitation by LSC (CompuCyte) (6). To show translocation, we programmed the LSC to measure the green fluorescence intensity in two separate cell compartments: 1) within the nucleus, as defined by the PI staining (integration contour), and 2) in the cytoplasm, defined outside the nucleus by a specified number of pixels (peripheral contour). Background fluorescence (background contour, outside cell membrane) was subtracted from both regions. A minimum of 2,000 cells were counted per condition, based on preliminary experiments showing statistical differences of defined scan areas containing ∼5,000 cells.
NF-κB activity ELISA. We determined the binding activity of NF-κB by a consensus oligonucleotide binding ELISA, using EZ-Detect Transcription Factor Kits for NF-κB p65 (Pierce), according to the manufacturer's instructions. Equal numbers of cells were cultured in flasks, treated with media or media containing 10 μg/ml LPS for 1 h, and then washed and lysed with NE-PER nuclear and cytoplasmic extraction reagents (Pierce), following the manufacturer's recommended protocol. The extracts were frozen at -80°C until use.
TNF-α measurement from BMC. BMC or AM/PM were cultured in respective incubators in 24-well plates and challenged for 4 h with either media alone or media with 10 μg/ml LPS. Culture supernatants were collected and assayed for TNF-α by ELISA kit (Pharmingen), against a standard curve according to the manufacturer's instructions. To ensure that differences seen were not due to unequal cell numbers after treatment, we repeated the experiments in eight-well chamber slides, and, after removal of supernatants for TNF-α assay, the cells were stained with PI (Molecular Probes). Identical-size scan areas were counted for multiple replicates of each treatment with the LSC. Cell numbers were shown not to vary significantly between different treatments (data not shown).
Statistical analysis. The statistical significance of differences between the BMC from two different oxygen tensions was determined by an unpaired two-tailed t-test with Prism statistical software. For statistical analysis of ratios, P values were calculated by the Mann-Whitney nonparametric t-test, with Prism software. We analyzed flow cytometric data using Kolmogorov-Smirnov statistics with the CellQuest software (Becton-Dickinson). A P value of ≤0.05 was considered significant and is represented as an asterisk. Error bars represent SE of a minimum of three replicate samples. Where data are presented in the text, the symbol ± indicates SE. Experiments were repeated at least twice with similar results, and representative data are shown.
Effect of oxygen tension on phagocytosis and APC activity. To demonstrate functional differences between the bone marrow cells differentiated in high- vs. low-oxygen conditions, phagocytosis and antigen presentation were assayed for BMC as well as AM and PM. The phagocytic capacity of PM compared with AM for E. coli uptake is shown in Fig. 2A. The fluorescence of PM following incubation with fluorescent bacterial particles was significantly higher than that of AM. Phagocytosis by BMClow was similarly increased over that of BMChigh in this assay system (Fig. 2B).
To measure antigen presentation activity by AM and PM, and similarly by BMChigh and BMClow, we cocultured the macrophages with T cells from the spleens of DO11.10 mice in the presence of OVA and assayed APC-induced T cell proliferation by BrdU uptake. DO11.10 BALB/c mice express a transgenic TCR that is specific for OVA presented on MHC II so that the T cells will respond to APC presenting OVA peptides, by proliferation and production of cytokines. As expected, PM induced significantly more T cell proliferation than AM (Fig. 3A). Figure 3B shows that BMClow were also significantly better APC in terms of inducing T cell proliferation than the BMChigh. As another measure of the macrophages' ability to activate the T cells, the supernatants from these experiments were assayed for IFN-γ by ELISA. There was significantly more IFN-γ released from T cells cultured with BMClow (14.9 μg/ml ± 0.63, n = 4) compared with those cultured with BMChigh (7.98 μg/ml ± 0.49, P < 0.05 compared with BMClow). Together, these results demonstrate that in functional assays, BMChigh performed similarly to AM, and BMClow behaved more like PM.
Effect of oxygen tension on intracellular oxidation conditions. To assess how the cultures' oxygen tensions affected intracellular redox state, dihydroxy-dichloro-fluorescein diacetate was added to cultures. After incubation with the reduced diacetate form, H2DCFDA, the fluorescence intensity of oxidized dye was measured in BMC from high- vs. low-oxygen environments. Figure 4 shows that AM and BMChigh had a significantly higher overall mean fluorescence intensity than PM and BMClow, respectively, suggesting the presence of a more oxidative state. Despite the apparently higher oxidative condition in BMChigh, measurement of intracellular GSH levels demonstrated an upregulation of GSH in BMChigh (2.3 ± 0.16 nmol/106 cells vs. 1.3 ± 0.1 nmol/106 cells in BMClow, n = 6 separate samples for each condition, P < 0.05 compared with BMChigh). This was shown to be consistent with higher GSH levels in AM compared with PM (1.6 ± 0.4 nmol/106 cells vs. 1.0 ± 0.1 nmol/106 cells, n = 6, P < 0.05) and suggests a possible compensation for the more oxidative environment. However, when the ratios of GSSG/GSH were calculated, AM and BMChigh both showed higher ratios (Fig. 5, A and B), indicating a more oxidized intracellular environment, consistent with the DCFDA data. These data suggest that the BMC model system induced similar redox conditions in cells as those measured in AM and PM. However, unlike AM and PM, the functional differences between BMChigh and BMClow can be attributed to the redox environment, since otherwise their culture conditions were identical. Consequently, the data strongly suggest that the single variable of the redox environment is responsible for these differences in cell function of the BMC.
Effect of oxygen tension on NF-κB translocation and TNF-α production. Intracellular redox state can affect cell function through modification of the activity of redox-sensitive transcription factors. To determine whether the redox environment was affecting NF-κB activity, we measured LPS-induced translocation of NF-κB. We performed immunofluorescent labeling of the NF-κB p65 subunit using antibodies to p65 in BMC cultured in the high- and low-oxygen incubators. Its translocation to the nucleus following LPS challenge (10 μg/ml for 1 h) was quantified by an LSC. This instrument allowed visualization of the localization of fluorescent staining, as well as quantification of the fluorescence in the nucleus vs. the cytoplasm (6). Figure 6A shows fluorescent images of the BMC with and without LPS challenge. In both BMChigh (panels 1 and 2, from the left) and BMClow (panels 3 and 4), the green staining around the red nuclei shows cytoplasmic localization of NF-κB in the resting cells (panels 1 and 3). After LPS challenge, the nuclei for both cell types appear more orange/yellow, demonstrating overlap of the green and red staining and suggesting translocation of NF-κB to the nuclei (panels 2 and 4). However, the translocation is more striking for BMClow in which nearly all of the NF-κB appears in the nuclei (panel 4). Figure 6B shows the LSC fluorescence data from replicate wells of the images shown, as ratios of nuclear- to-cytoplasmic staining. The data confirm that the translocation of NF-κB into the nucleus is greater in BMC cultured in low vs. high oxygen following LPS challenge and suggest that the NF-κB pathway is more active in cells grown in low oxygen. Similar to these results with BMC, there was a significantly higher ratio of nuclear to cytoplasmic staining following LPS challenge of PM than with AM (Fig. 6B), confirming more NF-κB translocation in cells from lower oxygen environments. In a separate experiment, the mean nuclear staining of NF-κB (by LSC) in LPS-treated cells increased 39% after 4 h in BMC from the low oxygen condition but was not significantly increased in the BMC from high oxygen (not shown).
To test the binding of NF-κB to its target DNA sequence, we added nuclear extracts of the treated and untreated cells to microtiter plates coated with NF-κB consensus oligonucleotides. The extent of binding to the wells by the NF-κB in the cell lysates was determined as luminescence using a fluorescence plate reader after addition of enzyme conjugated anti-NF-κB antibodies and enzyme substrate. Figure 6C shows increased NF-κB activity in the BMClow following treatment with LPS, whereas the BMChigh did not show increased binding with LPS treatment. These results further support increased NF-κB activity in the cells from lower oxygen environments.
Consistent with this activation of NF-κB, BMClow also produced significantly more TNF-α (a NF-κB-regulated proinflammatory cytokine) in response to LPS than BMChigh (Fig. 7B). Figure 7A shows a similar experiment using AM vs. PM, in which LPS-stimulated TNF-α production was significantly higher from PM than from AM.
The overall hypothesis for this study was that the normal phenotype of AM is in part due to the redox environment of the lung, which is relatively hyperoxic compared with other tissues. To test this hypothesis, BMC were differentiated under conditions where the oxygen environment could be altered. An oxygen environment similar to alveolar Po2 not only altered phenotypic parameters in a predictable fashion, but the resulting macrophages were similar to AM in terms of the functions tested. Both AM and BMChigh showed reduced phagocytosis of fluorescent particles compared with PM or BMClow, respectively. Reports are inconsistent regarding relative phagocytic capacity of AM compared with other macrophages, and this may be due to length of culture time, culture conditions, type of particle and opsonization (13, 26). Under the conditions used for this study, the macrophages from low-oxygen environments phagocytosed more bacterial particles, based on fluorescence intensity, than those from high oxygen. However, microscopic visualization of the cells showed that AM had taken up large numbers of particles, consistent with their function in clearing particles from the lung. Nevertheless, the functional similarities between the cells from different oxygen tensions in this specific assay were consistent with the hypothesis that oxygen tension could affect phenotype. The BMChigh were also less active as APCs than BMClow, similar to the relatively low APC activity of AM. These results suggest that BMC differentiated in the high oxygen incubator are similar to AM in terms of phagocytic capacity and antigen presentation, whereas BMC differentiated in low oxygen are more active phagocytes and APCs, similar to PM. Although the environments of AM and PM have differences besides oxygen tension, many of which could affect cell function, the BMC system allowed the examination of the effects of oxygen tension alone. The effects of different redox environments on these macrophage functions provide compelling evidence that redox regulation may be very important in contributing to the AM phenotype.
To understand the intracellular events regulating signal transduction in these conditions, it was important to determine how the extracellular oxygen environment affected the cells intracellularly and to demonstrate similarities between the two models, BMC and AM/PM. Earlier studies have shown that AM and PM have different levels of glycolytic and oxidative phosphorylation activities and that these metabolic differences not only are attributed to the different oxygen tensions but are also adaptive in that AM are more efficient in oxidative metabolism (22). BMC grown at 5 and 20% oxygen, similar to our model, developed metabolic characteristics that matched PM and AM, respectively (1). These studies suggest that environmental oxygen tension does modify the intracellular redox state and that this can lead to altered cell function. The experiments reported here provide additional details about the specific changes occurring and how those changes relate to other cell functions.
Despite higher levels of total GSH, both the DCFDA assay and measurement of GSSG/GSH ratios suggested that BMChigh and AM are both more oxidative intracellularly than their counterparts from low oxygen, possibly due to an inability of the cells to completely buffer against their environment. Murata et al. (18) have demonstrated that intracellular redox status can modify cytokine production by macrophages when challenged by IFN-γ and that cells that are more oxidized (less GSH) tend to produce less of the proinflammatory T helper (Th) 1 cytokine IL-12 than cells with more reduced GSH. This is consistent with our model in which cells with more reduced GSH appear more active in several macrophage activities. Because GSH can affect signal transduction through redoxsensitive transcription factors (7, 11), these data suggested that the redox environment might affect phenotype through NF-κB, a redox-sensitive transcription factor involved in many proinflammatory activities.
To explore a possible mechanism of the redox regulation of macrophage phenotypes, LPS-induced NF-κB translocation was analyzed. The LSC allowed visualization of the localization of the NF-κB p65 subunit by fluorescent staining, as well as quantification of the fluorescence in the nucleus vs. the cytoplasm (6). The translocation of NF-κB into the nucleus was much more evident in BMC cultured in low vs. high oxygen following LPS challenge, suggesting that the NF-κB pathway is more active in cells grown in low oxygen. The activity of NF-κB also appeared greater in the BMClow, based on a consensus oligonucleotide binding assay. To support these data with an effect on protein expression, we measured TNF-α as an important macrophage proinflammatory cytokine that is, in part, regulated by NF-κB. Consistent with the NF-κB results, BMClow produced more TNF-α following LPS challenge than BMChigh, suggesting that the redox regulation of NF-κB translocation led to downstream effects on proinflammatory cytokine production. Because NF-κB has been shown to coordinately regulate many functions of bone marrow-derived DC (27), this regulation of NF-κB could significantly impact overall phenotype of macrophages as well.
Overall, this study demonstrates a clear phenotypic difference between BMC differentiated in environments of different oxygen tensions, suggesting that at least some macrophage activities may be significantly regulated by redox mechanisms in vivo. GSH is a critical cellular redox buffer that has also been shown to regulate signal transduction through reduction of thiol groups in active sites of signaling components (7, 11, 20). Cells residing in the lung, where the Po2 is relatively high, would be expected to upregulate levels of GSH to provide sufficient buffering. Nevertheless, these cells would be challenged by their environment to maintain the GSH in a reduced form. The data presented here show that modifying extracellular oxygen tension does lead to more oxidized glutathione (GSSG), represented as increased ratios of GSSG to GSH, despite higher overall levels of total GSH. Because oxidative stress has been implicated as an activator of NF-κB (16), our data would appear to contradict this classic picture of NF-κB in that it was more active in cells with more reduced GSH. However, recent studies have challenged the validity of a causal relationship between oxidative stress and NF-κB activation (4). Oxidative conditions have actually been shown to inhibit cytokine-induced activation of NF-κB through oxidative inactivation IκB kinase (12). In the study reported here, the mildly oxidative conditions appear to have inhibited LPS-induced NF-κB activation. The mechanism of this unique redox setting is not yet clear, but other important redox buffers may be involved, such as thioredoxin, which can also affect NF-κB activation (8). In the lung environment, a mechanism must exist to protect the cells from oxidative stress without the often concurrent upregulation of proinflammatory responses, since that would lead to chronic lung inflammation. The data presented here support the hypothesis that AM have a unique redox setting that allows adequate oxygen buffering while downregulating proinflammatory activities. Disruption of this balance occurs following environmental exposures, often leading to chronic inflammatory conditions. A clearer understanding of the redox regulation of the AM phenotype could lead to improvements in the prevention and treatment of chronic inflammatory conditions.
Several interesting studies have recently addressed the possibility that intracellular redox conditions may play a role in skewing the Th1/Th2 immune responses (7, 17, 18). This study is consistent with several of these findings. First, a Th2 response has historically been considered “immunosuppressive” due to its ability to downregulate inflammatory responses. One of its landmark cytokines, IL-10, has been associated with downregulation of APC activity and T cell response in the lung (9, 19). Nevertheless, excessive Th2 responsiveness has been implicated in asthmatic conditions, suggesting a carefully regulated balance. In addition, measurement of IFN-γ in the supernatants from the APC experiments reveals an increase in this Th1 cytokine that was significantly higher when BMClow were used as the APC, suggesting that BMChigh might favor a more Th2 response. The ability of redox signals to skew the T cell response toward Th2 in the lung would contribute to its immune suppression. An understanding of this tendency in the normal lung could provide important clues to controlling specific pathological changes following exposures that disrupt this balance and explain the unique immune regulation of these critical players in lung immune homeostasis.
Although it has become evident in recent years that redox control of cellular function is of critical importance, the complexity of cellular oxidation/reduction regulation has prevented a clear understanding of how moderate changes in redox potential differ in their effects from those of the more extensively studied oxygen radicals in stress responses. This study explored these more subtle changes that affect phenotype. The AM and their role in lung homeostasis provide a unique model of cells exposed to high oxygen tensions that are also critical in protecting the health of the lung. The implications of redox control of AM phenotypes are profound, so that studies in this area may lead to much greater understanding of not only normal homeostasis but also the mechanisms that prevent a return to homeostasis in diseases manifested by chronic inflammation.
We gratefully acknowledge Dr. Terrance Kavanagh, University of Washington, and Dr. Jay Patel, University of Florida School of Medicine, for review and advice regarding this project.
This work was supported by Environmental Protection Agency Center Grant R828602 and National Institute of Environmental Health Sciences National Research Service Award Post-Doctoral Fellowship Grant ES-11249.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2004 the American Physiological Society