Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Chronic inflammation occurs when mononuclear phagocytes infiltrate into tissue, attracted by chemotactic molecules secreted by tissue resident cells (1). Infiltrated monocytes are capable of releasing other molecules that amplify chronic inflammation through direct injury as well as by attracting and activating other inflammatory cells. Monocyte chemotactic protein-1 (MCP-1) is an 8.7-kD molecule with potent monocyte-attracting and -activating properties (2). MCP-1, the human homologue of the mouse JE competence factor, is a member of the C-C chemokine family, which also includes RANTES and macrophage inflammatory protein-1 (MIP-1) (3). MCP-1 is produced by a variety of cells, including monocytes/macrophages, fibroblasts, and epithelial and endothelial cells. Although the role of MCP-1 in human disease remains to be fully defined, this molecule has a potentially important role in delayed hypersensitivity reactions. MCP-1 is most likely important in attracting monocytes to the lung where they mature into alveolar macrophages (4). MCP-1 has been shown to prime macrophages for tumor killing (5), suggesting this cytokine can also induce release of tissue-damaging molecules by mononuclear cells.

A number of stimuli, including lipopolysaccharide, phytohemagglutinin, interleukin-1, IgG aggregates, tumor necrosis factor, and others have been described for MCP-1 (2). In one study (6), oxygen radicals, generated in mouse mesangial cells after exposure to tumor necrosis factor or IgG aggregates, modulated MCP-1 expression through direct transcriptional activation of the MCP-1 gene. In that same study, exogenous synthetic reactive oxygen species, produced using xanthine oxidase and xanthine, caused a marked increased expression of MCP-1 mRNA also through enhanced transcription. In another study, enhanced MCP-1 expression was noted in rat kidneys exposed to ischemia followed by reperfusion (7), a situation in which natural xanthine/xanthine oxidase generation of reactive oxygen species occurs.

Production of another C-C chemokine, MIP-1, is also controlled by reactive oxygen species. Exposure to hydrogen peroxide and menadione (2-methyl-1,4-naphthoquinone), two oxidants, enhances MIP-1 mRNA synthesis by a rat alveolar macrophage cell line (8). Hydrogen peroxide stimulates MIP-1 mRNA production through direct effects on gene transcription as well as through stabilization of the MIP-1 transcript stability. There is also other information that suggests reactive oxygen species alter stability of a non-cytokine gene transcript. Hyperoxia, a condition that results in increased oxidative stress, induces a rise in lung catalase mRNA in rats through stabilization of the catalase transcript (9). These studies suggest oxygen radicals, including those generated by hyperoxic conditions, can modulate protein production through effects on RNA stability.

Pulmonary oxygen toxicity is an important clinical problem that occurs in patients receiving high inspired oxygen tensions for prolonged periods of time (10). Lung damage in this setting is due to increased formation of reactive oxygen species which can directly injure tissue. Amplification of the injury through stimulation of inflammatory cell influx by oxidants is also a factor. The histologic picture of hyperoxic lung injury consists of diffuse alveolar damage. Early in the temporal sequence of this reaction, there is an influx of polymorphonuclear leukocytes (PMN). Later, 72 h into the reaction in animals, there are increased numbers of macrophages (11). Increased levels of MCP-1 induced by hyperoxia could explain these increased numbers of macrophages because of the potential for this cytokine to attract to the lung circulating monocytes, precursor cells for macrophages.

We (12) and others (13) have previously demonstrated that hyperoxia stimulates production of interleukin-8, a powerful PMN chemoattractant that is in the C-X-C chemo-kine family. The purpose of the current study was to evaluate the effects of high concentrations of ambient oxygen on MCP-1 mRNA and protein levels. We also examined effects of high oxygen concentrations on MCP-1 mRNA stability. Our studies suggest hyperoxia enhances MCP-1 production, in part through effects on MCP-1 mRNA stability, and dexamethasone partially attenuates these effects.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Reagents and Materials

Hanks' balanced salt solution, modified Hanks' solution (without calcium or magnesium), and reagents for phosphate-buffered saline (PBS) with or without 2% Tween 20 were purchased from GIBCO-BRL (Gaithersburg, MD). U937 cells were obtained from American Type Culture Collection (Rockville, MD) and maintained in cell culture with RPMI plus penicillin/streptomycin and 10% fetal calf serum (FCS) (GIBCO-BRL) in the presence of 5% CO₂. Dexamethasone and actinomycin D were obtained from Sigma Chemical Co. (St. Louis, MO).

Hyperoxia Exposure

U937 cell cultures were performed in polystyrene tissue culture flasks. Cells were diluted in RPMI with 10% FCS for the planned assays. Cells were exposed to hyperoxia (95% O₂, 5% CO₂) or normoxia (room air/5% CO₂) in an oxygen chamber at 37°C for variable amounts of time in the presence or absence of dexamethasone (final concentration of 100 µM) as previously described (12). The oxygen chamber was constructed using Plexiglas and an airtight door. An inlet port was drilled for oxygen administration, and an outlet port was connected to tubing placed on water seal to avoid excessive pressure buildup. The gas mixture was infused until appropriate oxygen concentration, as determined by an oxygen sensor in the chamber, was reached, then gas was continued at low flow to maintain the appropriate oxygen concentration. Humidity was provided by placing a beaker of distilled water in the chamber. During all exposures, pH of the cell suspensions, as determined by phenol red color change, was physiologic throughout the time period. A previous study (12) demonstrated that exposure of U937 cells or human alveolar macrophages to this level of hyperoxia for 24 or 48 h did not significantly affect cell viability assessed by lactate dehydrogenase (LDH) release. After exposure, supernatant was separated from U937 cells by centrifugation and frozen at 80°C prior to determination of MCP-1 concentration by enzyme-linked immunoassay. Cell pellets were used for total RNA isolation and Northern blot analysis.

MCP-1 Measurement

An enzyme-linked immunoassay was developed to measure concentrations of MCP-1 in cell culture supernatants. Microtiter plates were coated overnight with 200 µl of a 1:2,000 dilution of a polyclonal antibody (Genzyme, Cambridge, MA) directed at MCP-1. Wells were washed and then cell culture supernatant or known concentrations of recombinant MCP-1 (R&D Systems, Minneapolis, MN) diluted in PBS with 0.05% Tween and 0.1% bovine serum albumin were added. Plates were incubated at 37°C for 1.5 h then wells were washed and 200 µl of a 1:1,000 dilution of a secondary polyclonal antibody generated in goats (R&D Systems), directed at MCP-1, was added. Plates were incubated for 1.5 h at 37°C and then washed and 200 µl of a 1:10,000 dilution of a horse radish peroxidase-labeled antibody, directed at goat IgG (Calbiochem, San Diego, CA), was added. Plates were incubated at 37°C for 1.5 h and then washed and a peroxidase substrate was added. The MCP-1 concentration was directly proportional to the absorbance at 490 nm and the concentration of this cytokine in fluid was calculated by comparing to the standard curve generated using recombinant MCP-1 standard. Results are expressed as nanograms per milliliter of supernatant.

Isolation of Total RNA and Northern Blot Analysis

After exposures and separation from supernatant, U937 cells were lysed with guanidine isothiocyanate. Cellular lysate was layered over CsCl (5.7 M CsCl, 0.1 M EDTA) and centrifuged in a Beckman SW-41 rotor at 35,000 rpm overnight. The RNA pellet was resuspended in TES (10 mM tris[hydroxymethyl]-aminomethane, 5 mM EDTA, 1% sodium dodecyl sulfate [SDS]) and extracted with phenol/ chloroform followed by chloroform alone. Total RNA was then precipitated by adding 3 M sodium acetate and absolute ethanol, pelleted in a microcentrifuge and dried prior to resuspension in RNAse-free water. Thirty micrograms of total RNA, suspended in a solution containing dimethyl sulfoxide, glyoxal, and 0.5 M sodium phosphate (pH 7.0), were loaded onto a 1% agarose gel (in 0.01 M sodium phosphate) and electrophoresed for 4 to 6 h at 60 V. RNA markers were run in parallel to establish size of detected transcripts in sample. Material on gel was transferred to nylon membranes by capillary action and was probed with cDNAs to MCP-1 (clone #65932; ATCC) and gamma-actin (clone #HFBCE81; ATCC) labeled with ³²P using random primers and purifed using a G50 push column. Labeled probes were added to achieve a concentration of 1.5 × 10⁶ cpm/ml. Membranes were first prehybridized at 65°C in a solution containing 50 mM NaPO₄ (pH 6.5), 100 mM NaCl, 50 mM PIPES (pH 6.8), 1 mM EDTA, and 5% SDS, and then the probe was added and membranes were hybridized in the same solution overnight at 65°C. The next day membranes were washed at 65°C four times in 5% SDS, 0.667× SSC (3M sodium chloride, 0.3 M sodium citrate, pH 7.0), and then membranes were exposed to film overnight. Densities of bands noted on film were quantitated by densitometry. Results are reported as percentage of control condition as previously described (12). Statistical analysis was performed using raw data so that normoxia effects could be properly compared with the other two treatments.

RNA Stability Studies

In certain experiments, actinomycin D (final concentration of 10 µg/ml) was added at the onset of cell culture to stop gene transcription (14). Cells were removed prior to actinomycin addition as a zero time control, and then cells were exposed to hyperoxia or normoxia and aliquots were removed for total RNA isolation at 2, 4, and 6 h after beginning of exposure. In another set of experiments, cells were exposed to hyperoxia with or without 100 µM dexamethasone after actinomycin addition and RNA was isolated 2, 4, and 6 h after start of exposure. Northern analysis was performed using the MCP-1 and gamma actin cDNA probes. The MCP-1 signal, quantitated by densitometry, was divided by the actin signal at various time points after exposure and was expressed as the percentage of the zero hour control. Effects of hyperoxia or dexamethasone treatment on MCP RNA stability were determined by comparing with the appropriate control.

Statistics

Data were stored and analyzed using a software package (PCSTAT; Statsoft, Tulsa, OK) and an Insight 486 computer. Differences in means were assessed by paired Student's t test. When multiple groups were involved, differences were assessed by ANOVA and individual group differences were assessed by the Newman-Keuls test. All data represented are mean ± SEM. P values  0.05 were considered significant (15).

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Stimulation of MCP-1 Production by Hyperoxia and Effects of Dexamethasone

U937 cells exposed to 95% oxygen for 6 and 24 h produced significantly more MCP-1 than did cells incubated in normoxia (Figure 1). Although cells exposed to hyperoxia for 2 and 4 h also tended to produce more MCP-1, these differences were not significant. Addition of dexamethasone at the initiation of exposure significantly attenuated hyperoxia stimulation of MCP-1 production 4 and 6 h after onset of hyperoxia.

View larger version (24K): [in this window] [in a new window]   Figure 1.   Effects of hyperoxia exposure in the presence (gray bars) or absence (black bars) of dexamethasone on U937 production of MCP-1, determined by enzyme-linked immunosorbent assay. White bars show MCP-1 production in the presence of normoxia. (n = 5, *P < 0.05 versus normoxia control and hyperoxia in the presence of dexamethasone.)

Enhancement of MCP-1 mRNA Expression by Hyperoxia and Effects of Dexamethasone

U937 cells cultured in normoxia for 2, 4, and 6 h did not demonstrate any significant difference in MCP-1 mRNA levels. However, when U937 cells were exposed to hyperoxia for 2, 4, or 6 h, there was a significant increase in MCP-1 mRNA levels (normalized to actin mRNA levels) 6 h after initiation of exposure when compared with normoxia control (Figure 2). Dexamethasone added at the beginning of the exposure attenuated this increase. The effect of dexamethasone increased over time; the maximal effect noted was after 6 h of exposure.

View larger version (27K): [in this window] [in a new window]   Figure 2.   Effects of hyperoxia exposure on U937 MCP-1 mRNA levels, expressed as the percentage of normoxia control. The graph shows densitometry data for five separate experiments (MCP signal/gamma actin signal [mean ± SEM]: white bars, normoxia; black bars, hyperoxia; gray bars, hyperoxia + dexamethasone) expressed as the percentage of the normoxia control, which is set at 100%. The autoradiograms show representative Northern blots of mRNA from U937 cells exposed to normoxia (RA) or hyperoxia in the absence (H) or presence (H+Dex) of dexamethasone. The Northern blot was probed with cDNA to MCP and gamma actin as described in the text. (*P < 0.05 versus normoxia control and hyperoxia with dexamethasone.)

Alteration of MCP-1 mRNA Stability by Hyperoxia and Dexamethasone

When transcription was halted by actinomycin at the beginning of exposure, the decrement in MCP-1 transcript was slower in hyperoxia-exposed cells than those in normoxia (Figure 3). U937 cells exposed to hyperoxia in the presence of dexamethasone exhibited diminished MCP-1 mRNA stability when compared with hyperoxia alone (Figure 4).

View larger version (41K): [in this window] [in a new window]   Figure 3.   Effects of hyperoxia on MCP-1 mRNA stability. The graph shows MCP-1 mRNA signal at various time points after exposure of U937 cells to hyperoxia (gray bars) or normoxia (black bars). MCP-1 signal (normalized to gamma actin signal) was detected by Northern blotting. The MCP/actin ratio in cells harvested prior to actinomycin exposure (zero time) is expressed as 100% and subsequent time points after exposure to the two conditions are expressed as a percentage of that value. The autoradiogram shows a representative Northern blot probed with MCP-1 or gamma actin cDNA. (n = 5, *P < 0.05 versus normoxia.)

View larger version (33K): [in this window] [in a new window]   Figure 4.   Effect of dexamethasone on MCP-1 mRNA stability. The graph shows MCP-1 mRNA signal (normalized to the actin signal) at various time points after exposure of U937 cells to hyperoxia in the absence (black bars) or presence (white bars) of dexamethasone. The autoradiogram shows a representative Northern blot probed with MCP-1 or gamma actin cDNA. (Mean ± SEM, n = 3, *P < 0.05 versus hyperoxia alone.)

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

We have demonstrated that U937 cells produce MCP-1 and production of this cytokine is increased by exposure to hyperoxia. In addition, dexamethasone attenuates the effects of hyperoxia on MCP-1 production by U937 cells. Hyperoxia-modulated increase in U937 cell MCP-1 production is associated with an increase in MCP-1 mRNA expression which is due in part to an alteration in MCP-1 mRNA stability. Dexamethasone pretreatment attenuates the rise in MCP-1 mRNA levels mediated by hyperoxia, also partially through effects on mRNA stability. This is not related to effects of hyperoxia on cellular viability, as we have previously demonstrated that LDH release is no different in cells cultured for 24 and 48 h in the presence of hyperoxia as compared with normoxic conditions (12).

We have demonstrated similar effects of hyperoxia on U937 and macrophage interleukin-8 production in the past (12). Effects of hyperoxia on interleukin-8 mRNA stability were not examined in that study. Other investigators (6, 7) have demonstrated that reactive oxygen species can modulate MCP-1 production, although effects of hyperoxia have not been studied. Satriano and colleagues (6) showed that reactive oxygen species generated by hypoxanthine in combination with xanthine oxidase upregulate MCP-1 mRNA expression in mouse mesangial cells. Their studies suggested MCP-1 mRNA upregulation was due mostly to transcriptional activation of the MCP-1 gene by reactive oxygen species. However, effects on MCP-1 mRNA stability were not studied. Our study suggests oxidants also alter MCP-1 mRNA degradation, probably due to inactivation of specific ribonucleases. The gene coding for MCP-1 contains AUUUA-rich sequences in its 3' untranslated region which are associated with mRNA instability (16). The current study suggests normal MCP-1 mRNA instability can be partially reversed by exposure to hyperoxia, resulting in increased MCP-1 mRNA and protein levels. This may be due to susceptibility of specific ribonucleases to oxidation.

There are other mechanisms by which oxidants generated by high inspired oxygen concentrations or oxidant-producing enzymes might modulate MCP-1 production. As noted, previous investigators have demonstrated that reactive oxygen species activate MCP-1 gene transcription. This effect may occur either through a specific oxidant-responsive element on the MCP-1 gene or through oxidant effects on the transcriptional regulatory proteins, such as nuclear factor kappa B (NFB), which has been shown to be regulated by an oxidizing environment (17). Sequencing of the gene for MCP-1 has shown a putative binding site for NFB (18). Our studies suggest that oxidants at least partially exert effects on MCP-1 production through modulation of MCP-1 mRNA stability. However, direct stimulation of MCP-1 gene transcription could still be a factor in regulation of MCP-1 production by hyperoxia.

Our study also demonstrates that dexamethasone attenuates hyperoxia-stimulated enhanced MCP production. This effect is due in part, but not totally, to effects of dexamethasone on MCP mRNA stability. Other investigators (19) have shown that the mouse competence factor JE gene transcript is degraded more rapidly when cells are exposed to corticosteroids. JE shares 68% amino acid homology with MCP-1 (16). Corticosteroids have also been shown to affect stability of the interleukin-6 transcript in human fibroblasts (20). In addition to effects of cortico-steroids on MCP mRNA stability, these agents could have direct effects on gene transcription through interaction of the glucocorticoid receptor complex with a steroid-responsive element on the MCP-1 gene. In fact, transcription of the JE gene is directly inhibited by exposure to corticosteroids (21). This inhibition is reversed by exposure to cycloheximide, a protein synthesis inhibitor, suggesting JE/ MCP-1 gene transcription is directly inhibited by a protein induced by corticosteroids. Our data suggest stability of the MCP-1 transcript in the presence of hyperoxia is decreased by corticosteroids. The gradual increase in dexamethasone effect is consistent with stimulation of a negative regulatory process such as induction of specific ribonucleases. The degree to which dexamethasone affected MCP-1 mRNA stability in our study does not appear to totally explain the effects of dexamethasone on MCP-1 mRNA and protein levels. Dexamethasone most likely affects MCP-1 production at both a transcriptional and post-transcriptional level.

In the current study, we used U937 cells to study regulation of MCP-1 production. Characterization of this histocytic cell line that was first isolated by Sundstrom and Nilsson (22) shows these cells to have characteristics of monocytes and macrophages. To our knowledge, the current study is the first to report that U937 cells produce MCP-1. This finding is not surprising and it may prove useful for the future study of this cytokine. Although U937 cells should not be considered normal, they are easily obtained and accessible to manipulations. Our study demonstrates that hyperoxia stimulates MCP RNA production by these cells 6 h after onset of exposure but the protein product is not significantly increased until after 24 h of exposure. This finding is consistent with findings by other investigators who have found that secretion of increased concentrations of MCP may lag, by several hours, increased MCP RNA levels. In one study (23), interferon- maximally increased MCP RNA levels after 3 h of exposure but maximal stimulation of MCP secretion was not noted until 24 h.

The role for MCP-1 in oxygen toxicity is undefined. The histopathology of hyperoxic lung injury consists of diffuse alveolar damage with desquamation of epithelial cells into alveoli, acute parenchymal inflammation, and proteinaceous material filling airspaces (10). We (12) and others (13) have previously demonstrated that hyperoxia stimulates interleukin-8 production by monocyte lineage cells. This may explain the intense acute inflammation that occurs early during hyperoxic damage. The current study suggests a mechanism for accumulation and activation of macrophages in the lung after hyperoxic exposure. This may be a mechanism for prolongation of the lung injury caused by high concentrations of inspired oxygen. Oxygen toxicity is associated with histologic evidence of increased numbers of alveolar macrophages derived in part from circulating monocytes attracted to the lung by cytokines such as MCP-1 (5). These macrophages have the potential for producing tissue-injurious enzymes as well as molecules that can attract more peripheral inflammatory cells, resulting in further amplification of the process.

Our study suggests high inspired concentrations of oxygen stimulate MCP-1 production. New properties and functions of MCP-1 will certainly be discovered in the future. Our study and studies by other investigators suggest production of this chemokine is controlled by oxidants. This finding may eventually be important in designing strategies to prevent acute lung injury due to increased concentrations of reactive oxygen species from progressing into irreversible injury.